Serious problem with XML2 current. Is this happening to you???

The main reason I ask this is to see if anyone else is getting similar results.

Ever since XML2 was first released I have been able to get around 4 amps of current (or more) from nearly every emitter. I've had the occasional delinquent VF like everyone but mostly results were consistent.

Then along came U3 bin. Almost every one of these I ever got had disastrous VF. Using exactly the same setup (26650 in my shorty mags) I would get as low as 2 amps often, up to maybe 3.5 or 3.7A max on fresh cells.

With the introduction of U4 bin things went back the way they should be. 4- 4.5amps on a single cell no problem. When RMM bought his first reel of them I picked up a large amount of them and never had any problem.

Recently I order a couple more batches from another source I find to be very reputable and all of them test nearly identical to what the U3s did. I did a controlled test of about 15 emitters to try and week out any anomalies, but sadly they all tested within .1V of each other. Even the very lowest VF ones in the batch were 4.1V @ only 4 amps :(

Is anyone else having this issue?

Just to be clear. Yes I have considered ambient temp and done additional testing inside at 70 deg. Yes, multiple identical setups in flashlights etc.

How about hitting one of them with 5-6A for a short time. Won’t that permanently lower the Vf of that emitter a little?

I bought a big batch of U4s before and I am still on it. I have not tried the newest U4s so I am not quite sure. Mines pull 4A+ no problem with a VTC5

Might they have sent you U3s or got a mislabeled batch from Cree?

You never did mention what cells you generally use.

Unless you're really lucky if you do that to a newer XM-L2 or even a bunch of the XP-L HIs you will raise the vF to at least 200-1000 volts, which is about how much I think it will take to jump the spark gap between the now vaporized bond wires.

Matt, 4.1V at 4A is crazy high for an XM-L2. I don't recall ever seeing one that high before. I check the vF before and after every single LED I dedome, so I see them several times a week and I don't think I've ever seen a batch of XM-L2 LEDs like that. The only 3V LEDs I have here that will go up to a true 4.1V without poofing are some of the newer XP-G2s which need that much voltage to hit 5A-6A.

Somebody design a driver that pushes amps no matter what the forward voltage. Threshold voltage? Pfffff! I’m a driver not a designer! Direct drive? Do you think I’m building it?

Laugh all you want, most of what we do is science fiction from a very short time ago. I hope it keeps happening :wink:

I ordered a few U3 1A from LED DNA, different months, batches and shipping dates - they are all fine.
on Samsung 25R, 30Q and other cells.


I hear you about the burn in. That is something we used to do with SST-90 emitters and it actually worked. I have attempted it with XML2s and it has no effect.

What RMM was saying about the spark gap thing is just because the new (U4 bin XML2s can only handle about 4.5A. Once it a while you will get all the way to 5, but I've not had any live beyond that.


What is the typical VF your getting with XML2s ( for example at 4 amp)? What change do you typically get after a de-dome?

Same question about XP-L and XPL-HI in top bins if you have tested any of these?

Thanks also to everyone else who has commented. Your input is valuable to me.

At least with my basic understanding of electronics, that constitutes a boost driver, and some sort of feedback to measure the Vf. Not exactly the easiest thing to get to work, much less at high currents.


Not trying to go off topic. What about total lumens and total wattage (VxA)?

Not off topic at all. Just now I did a test of the output as well. Last year I performed a series of controlled tests in order to calibrate my sphere and to determine the usefulness of the XHP-35. I happened to have a control test of one of RMMs original U4s which I had reflowed to a 1"x10" copper rod.

Lights used for calibration 8/10/15 Measured FC multiplier from FC
Shorty mag as tested by Ti-force 875 lumen 2710 0.322
4sevens X10 640 lumen 2000 0.32
4sevens Quark AA2 280 lumen 914 0.306
Test #2 Shorty mag 875 lumen 2710 0.322
Multiplier used for testing below .32 All emitters were reflowed to copper MCPCB and to solid round stock.
Cree XM-L2 U4 Measured Lumens Cree Chart
1000ma 432 459
1500ma 620 629
2000ma 793 824
2500ma 900
3000ma 1100 1122
3500ma 1238 -----------------------
4000ma 1369 -----------------------
4500ma 1478 -----------------------

Just now I re-calibrated my sphere to be sure, which as it turns out was within .0011 of where it was at before.

Cree XML2 U4 in question for VF and output 1/27/16 reflowed to copper rod as before
Calibration. Shorty mag .347 47s X10 .324 Quark AA2 .311 Average .327, multiplier used .33
1000ma 415 3.2v
2000ma 739 3.6v
3000ma 1032 3.9v
4000ma 1267 4.1v

Its impossible to tell for sure since I had only one test of the previous emiter, but it appears that the ones I have now track almost exactly one bin lower.

1a 432.............415
2a 793.............739
3a 1100...........1032
4a 1369...........1267

At this point I think more info is still needed to be sure this is the problem though since even emitters in the same bin can gap about that. RMM is going to send me some results from his testing later.

I am going to do some tests later tonight when I dedome some emitters and post the vF findings here of some of the XM-L2 U4 1C I have.

@vestureofblood: What are you using to measure current and vF?

Well thats strange, so far i have never ever met a single xml2 emitter that aint capable of running at least 4.5 amps easy,those u3 0d that i do love so much lately are running at 4.7-5amps on a fresh HE2 battery yet they are as good as U4 1C considering the lumen output, but the tint when dedomed is good

The best of those that i have found so far are both KD ans IOS V6 1A: easy hitting 6amps with a good 26650, a localy picked V6 2c-2d( same as above, IOS one aint that good ) , ofc V6 0D from IOS( they all hit 5.5amps, i bought 20 of those) and FT V6 0D-yet the tint here aint that good, prolly nto a true 0D but kinda 0C more or less

The one thing that is strange to me about V6 bins in thay they dont dedome that easy as the old ones

My bench power supply. I just took a short piece of heavy lamp cord (about 2ft) and wired it to the supply.

Is this the right way?

Beyond this I have been coming up short on current/lumens else where with these.

One example is an 18650 pocket light. Original tested at 1080 lumen @ 1 min. With this latest batch of emitters I was only able to get a very best of 960. And the 960 was after I over sized the wire, copper braided the tail spring, and progolded the clickie switch. All 3 of which I did not have to to in the roll out.

I do my testing with a similar branded bench supply. I dont use the readouts on the power supply itself but measre at the led as there is a slight difference between readings. The readings on my power supply are as accurate as my DMM's.

Measure voltage across the emitter with a DMM and compare that to what you get at your PS.

On a de-domed U4 1C in my modded Convoy L2, I measured 5.47A at the tail using my new clamp meter on a Basen 26650 4500 cell @4.22v, and DD FET driver.

On a modded BLF Lucky Sun D80 on a 30Q @4.21v: 5.15A tail (DMM), 5.36A tail (clamp), lumens: 1812 @start, 1744 @30 secs, 52.5 kcd (taken at 5m)

Don't have the exact #'s but on a XM-L2 T6 4C FET+1 driver Warsun X60 mod, I measured bout 6.5A tail on a DMM, and over 7A at the tail on a clamp meter. This is with an EFEST 4200 26650 or the new Basen 4500 26650, both at about 4.21v or 4.22v..

You should be getting between 5.0A and 5.5A on a U4 1C with a decent DD FET driver using a clamp meter (the UT-210E). As shown above, a clamp meter will always get a higher reading on a DD driver. On the high Vf U3's, I get in the range of 4.0A to 4.3A with a similar setup.

Your amps measured are consistently lower, so I'd suspect either the measuring method or higher resistances in the circuit (driver, host, wires, possibly MCPCB/LED, etc.).


Just now I did as you suggested and measured the VF with my DMM using the same test emitter as before. I got 3.99 @ 4 amps. The PS reads 4.1v so I am getting a slight variance there.


Thanks for chiming in.

Where you said

"You should be getting between 5.0A and 5.5A on a U4 1C with a decent DD FET driver using a clamp meter (the UT-210E). As shown above, a clamp meter will always get a higher reading on a DD driver. On the high Vf U3's, I get in the range of 4.0A to 4.3A with a similar setup."

That is fairly close to what I was getting wtih the U4s until recently. I dont have a clamp meter but I use heavy custom leads for my DMM. On a 26650 light I was averaging 4.4-4.8A on 26650 lights and just a little under that with 18650 lights.

These new ones wont squeak out anything above 3.8.