I think people are worrying too much about possible meter readings rather than how a light works for its purpose ie giving light in a dark place. I dont know about other people but that’s the important thing for me.
What I did notice is same cell and driver, an xm-l pulls 3.7a xm-l2 pulls 3.3a and looks brighter. I’m happy with that and will stop worrying.
I basically agree, but figure you are talking up to about $10 in parts difference (XML U3/basic star vs. XM-L2 U2/SinkPAD), and whether it looks brighter only for the first minute, then remaining time looks duller plus a shorter runtime - sure would be nice to know the trade-offs going in, pros and cons, so we can accurately answer these questions, like what Garry is asking here. Also for me, it affects your decision on amps and drivers, if 3.7A is unachievable, then 3.4A driver setup is fine - don't add extra 7135's or bother gping to a higher rated driver.
You are correct; one must consider the different junction temperatures at which the two emitters are spec’d. I made made this quick spreadsheet to illustrate ramifications. Keep in mind these are “typical” values. I wish Cree would disclose normal distribution data. If they did, the term “typical” would carry some meaning.
that was the part I left off, I was in a hurry.
what we do know is that a neutral xm-l2 will put out about the same lumens as a u2 @3a.
So I can now have a nice tint AND high output with lower running current - less load on the cell and switch components, cooler operating temperature.
Or put another way, you could now have an xm-l2 edc at for arguments sake 2.5a and have similar output to an xm-l at 3a (same bin emitter) again, less heat, so a longer run time on high.
I stand by my point, some people (none in this conversation I will add) are too busy worrying about the theory instead of looking at the practical reality, or playing to the new emitters strengths ie higher output from a nicer tint, that for me is the xm-l2’s forté. :bigsmile:
Exactly. The important part is that at worst, you can do the same with less with the XM-L2.
thumbs up
Plus OP meant without buck.
If you have a driver with boost, good luck haha, as it’s probably for 2 emitters in series and you would connect it to a single emitter, over voltage, emitter death.
No, he meant boost. If the cell is at 3.6V and you would like to have 3.8V at the LED, you need a boost circuit.
I did not mention numbers, was just trying to explain what is going on assuming the vf of the xml2 is higher than xml. And I must admit I am not sure where exactly idea came from, could very well be from match's retracted graph (@3A his graph gives a vf of 3.7V, where Cree claims are between 3.3 and 3.4V, that is significant different.)
Slim Pickens posted the vF numbers which are interesting, reversing the roles which is what was expected. What's also interesting is the lower the temp, the higher the vF which is what we see with SinkPAD's. Did an interesting test: C8 w/XM-L2 U2/SinkPAD 3.85A Nanjg, and C8 with XML U3/stock star 3.8A KD V2 - used same fresh charged LG 2400 mah batts from a Dell notebook pack pull. These batts seem to be average resistance (they take long to charge also).
C8 w/XM-L2 measured 3.4A, C8 w/U3 measured 3.8A. Simple wall and ceiling eyeball test (3 opinions) agreed the XM-L2 light was brighter. This match's what Gords is saying too.
I made a XM-L2 T6 3C drop-in with 11 380mA chips (4.18A) the best I can get is 3.86A with fresh off the charger unprotected Sanyo 2600 @4.20v but the current drops immediately and about 10 seconds later my battery is down to 4.17v and the current is down to about 3.7A
Guess I need a stout IMR/INR type cell.
BTY xm-l2 @ 3.8A is bright 
after about 10-15 minutes 3.2A-2.8A it’s still bright enough to light up the whole back yard.
I plan to do the same. Would like to compare with the driver I mention below. Would be cool do to same test with XM-L and XM-L2.
I used the XM-L2 (U2 KD claims on their drop-in, I don’t trust that store). Replaced the driver with the high output 3-mode Ultrafire driver. (Came with a popular P60 about a year ago)
Here is the data from the datasheets presented by emitter, and by temperature, i didn’t extrapolate beyond 3A, but someone else can feel
free to do so
I wonder what real world values will be like. I use the panasonic 3400s which don’t hold their voltages too well. I was planning on building an XML-2 drop in with sinkpad + linear driver which would only make this high Vf issue worse.
Do you guys think it would just be better to use the standard XML or maybe just not use a sinkpad?
I just added a feature to my LED analyzer device (Any interest in a LED/Battery analyzer device?) that lets it sweep the LED with a range of voltages and records the various LED parameters. He it is sweeping an XML from 0 to 1.8 amps. At an LED Vf of 3.000V the LED was drawing 1.225A and 3.675 watts. Don’t have any XML2’s to play with.
I’m considering using a sinkpad and XM-L2 in a host too small to dissipate all the heat, so being a bit warmer will lower the Vf, i haven’t done the math, but i am hoping for more output vs the stock XM-L on aluminum pcb
Looking at the first two graphs GB posted in #6, if you take into account the different scales at the bottom they seem almost identical. Both start and end at about the same place. To be usefull as a visual comparison, the scales need to be the same.
If you keep the XM-L2 around 2.5A you should be good to go. I look at it this way, you can get a nice neutral tint XM-L2 with about the same brightness as a XM-L U3 but the U3 will have a cool white tint.
I only use high mode for short 3 or 4 second bursts anyways, low and medium mode is what I use the most.
Exactly, me too - considering it for a Convoy M1 - got it running at 3.85A but with the stock U2 emitter and getting only 10% better over the 2.8A stock setup. Interesting because it gets hot quick and on the light meter in the lightbox, it dropped quick, real fast for the first 2 secs - seems like no where to go with the heat. Im suspecting maybe a bad LED reflow or bad star maybe because I got a good thermal grease under it, and sanded/polished the contact surfaces, but no room to add any copper in the pill - really like the quality of this small host though. Think a SinkPAD could be a big benefit for this type of small host design - will be doing it and checking the result.
funny, thats the light i am currently using
Anyone know if the LG 4.35v batteries up on FastTech today would help the Vf issue with XM-L2 single cell lights? I suppose it helps if you could charge them to 4.35v - I don't have a charger than can do that.
Another interesting battery today up on FT is: Sony US18650VTC3 18650 1600mAh 3.7V , rated at 30A discharge. If it's lower resistance than the 20A rated Samsung 20R's, it should do better on the XM-L2/SinkPAD high amp single cell lights.