Emitter Test Results pt4: XML2, XP-E2, XP-G2, and Nichia219! (along with sinkpad tests)

thanks for the data Match, it’s fantastic to see such controlled experiments especially as the data can then be compared back to previous emitters.

So the XM-L did ~875lm @3A, the XM-L2 (U2 I presume) does ~1000lm at 3A and the XM-L2 on a sinkpad does ~1100lm. 225lm extra for an emitter swap is pretty amazing! At 1.5A the figures are ~550, 630 and 650lm, which probably explains why I didn’t see much difference by eye when I swapped my XM-L T6 for XM-L2 T6 on Sinkpads.

It’s interesting that, despite the higher Vf, the XM-L2 on Sinkpad is more efficient at ~101lm/W @3A vs. 96lm/W on a plain aluminium star. Pretty damn amazing how far things have come in the last couple of years.

I just modded a KE-5 that did 4.8A (stock T6 on 16 mm alum star) - 920 lumens stock, to Nanjg @3.85A U3 1B/20mm SinkPAD and got 1,100 lumens. So, lower amps, cooler running, better output. At stock, that light got noticeably warm at 30 secs - after the mod I ran it to 6.5 mins where hottest spot got to 118F, pretty hot, but that's tailstanding on it's own @room temperature.

The XML2 is a U2 bin, and IIRC the XPE2 I sent was Q5 or what ever the top available bin for them is.

Yes, you are correct texaspyro. There is some loss because of the smaller amount of metal in a flashlight.

I think however you will be pleased to know I have conducted some testing on this. In the body of a small to medium size light ( like a 30mm head diameter) I put one of these XML2 on a sinkpad and wired it directly to my bench power supply. Even in the small light body XML2 sees gains at the 1 minute mark as high as 5800-6000ma. The output does have a gradual but steady decline at this current. Its more stable at slightly lower current like 5000.

I will be posting a thread reflecting some of the data in this experiment over the weekend.

great measurements!

If you put this combo in a well heatsinked flashlight that can handle the 5A (say, a few 18650's parallel, or one fatty INR), I wouldn't bother using a driver at all, just direct drive it on the batteries: 1600 lumens at 4.2V, still 1250 at 3.8V

Why is the forward voltage higher with XM-L2 on copper vs aluminum at the same amperage?

Bort: I think it’s because the emitter runs cooler on copper.

Also keep in mind that given two emitters of the same basic type, bin, and tint can still perform differently. Good 'ol cree tolerances.

correct

Tried Panasonic PD's in a couple of XM-L2 U2/SinkPAD lights and definitely as good or better than a AW IMR 2000. Also tested 2 Samsung INR 1500's from FT (cheap, but out of stock now), and also as good or better than the PD's. I was able to get 3.85A on my XinTD, and in the 7G9, I put 2 INR's and 1 PD (have only 2 each) and actually got 4.2A measured to match the 12 7135's, throw of 125 kcd measured! So the low resistance batteries do help big time. Working on a Small Sun ZY-T08 (2 parallel) now with XM-L2 U2 1C/SinkPAD Nanjg 4.2A - hoping it will be one hot setup thrower with the PD's or Samsung INR's.

Gentlemen,

The latest and final round of emitter tests are done and complete. The first post has been updated to include the Nichia219, along with the xpg2 and xpe2 comparisons of aluminum pcb vs the copper sinkpad.

Also, you'll notice that I deleted the vF graph for the XML2. For complete transparency reasons, here's why: During an equipment audit, I noticed that my trusty Fluke wasn't that trusty anymore despite being within it's cal date. I have since swapped it out with another proven unit and ran a retest. New vF graph to be up shortly.

-Match

nice,thanks again!
the xpg2 test is done with R5?

Thanks for the graphs! Sad to hear that you will retire.. and that there is no Nichia on copper test. ;)

Anyways, thank you. :)

Great work Match! Look at that XP-G2 go! Those are XM-L T6 numbers on a much smaller die. Explains the TN31mb throw results a bit.

That test helps make sense of the TN31mb. I’ve tested mine at 741otf at 30sec. I just wondered how in the crap in Saabluster do that. But now I can see a XP-g2 on copper can get those type of numbers. I wonder how much difference the dedome made on its output? Likely not much I would guess since it still has 741 lumens.
Anyway, thanks Match for the very informative tests. Awesome work!

Measured at the tail, my X7 driver pulls 4.59A from 3 x 18650 cells, so the emitter probably sees around 4A. Id like more throw so Im thinking about swapping the SST-50 for a dedomed XPG2 or XML2 on sinkpad. 4A would be pushing it for the XPG2 and Im not sure it would survive. The TN31 MB puts out a beam like a laser on dedomed XPG2 and Id like to have something similar. I have no idea how the X7 reflector would focus an XPG2/dedomed.

Theoretical:
XPG2 - 925 lumens before dedoming. After?… Im not sure. Maybe 600?
XML2 - 1390 lumens

Anyone have any thoughts on this?

Match, thanks for all the data and your hard work. The new possibilities are almost endless!

Match thank you very much for the testing & results. They are very usefull, appreciate it!!

Match, your nichia numbers reflect what I've seen on mine. I have a nichia 219 in a tr-801 with a driver that lets an xm-l pull 3A. With a nichia it pulls only 1.65A which coincides with your findings.

Again match your work is invaluable. Thanks for the huge effort doing this for us. My only concern though is that no one will build a light like we have before. I'm doing a mod on a particular flashlight at the moment that I thought was all planned out. XM-L2 on a copper star on a beefy copper pill in a large ally housing running a 5 amp buck boost driver. Do I now have to get a 9 amp driver to run the XM-L2 at its peak? Bugger, it never ends. Cheers and thanks again.

PS. Maybe I could use the 5 amp driver on a killer XP-G2 build?

wow… great data … very impressive results from the xpg2! it just about beats an xml when mounted on copper!

These types of post are one of the reasons I check forums. Provides such a wealth of information! You’ve inspired me to build an integrating sphere once I find the time.

1 for nichia on copper sinkpad Vf if possible!