When will Cree release xm-l3

So, you know that xm-l2 is quite old led(at least to me), there should be xm-l3 now
I know someone will say: there is already xp-l2 led, we dont need xm-l3, but
Xm-l led has bigger foorprint, bigger die, so the beam maybe flooder than xp-l led, bigger footprint means there is more area to reduce temperature
And the most important thing is maybe it can produce 1700+ lumen with low vf(we can already do with xp-l led, but the output is not regulated)

So
Cree
When will you give us xm-l3 led

Ps:sorry for any typo/grammar mistake
I’m not good at English, :frowning:

I think it will not come.
As you said yourself, we have XP-L2 now.
The trend is a small footprint.
5050 size is for XHP50 / XHP50.2
More power on smaller size.
Remember XR-E ? :wink:

If they can get a single die in 5050 size to have almost comparable output to XHP-50, it might be worth doing a XM-L3 for that. But yeah, I’m not gonna hold my breath!

Parallel connected XHP50.2 can be named XML3.
Not sure there is any need in parallel multi-die leds anywhere except of flashlights, I think in future we will have lots of high-voltage leds just because they are more usefull in industry lightening and etc.

But xhp50.2 only available in 6v and 12V( we can do the hack. But it’s very tricky)

I don’t care for parallel multi-dies. I want a single die rated for well over 2000+ lumens in a 5050 package size.

If they can make a single die led rated 2000 lumen max, that would be a very big step, because xm-l2 is rated only 1100 lumen( as far as i remember)

XP-L2 can do 1500 Lumens at 100 Lumens per Watt (15 Watts at 4.5 Amperes.)

Not really, because the L stands for the die size, the M stands for the 5050 footprint.
So the XHP50 could have been named XM-P or something like that.

Well, here is my logic:

  1. The XP-L2 is what I’d consider the latest die in this size (XPL die is about the same size as XML die). That has a top official rating over 1300lm on a much smaller package size. Smaller package means slower heat dissipation. Better thermal properties lead to higher output.
  2. The XHP-50 (both version one and version two) have a top rating over 2500lm in the same package size as XM-L2. They use smaller dies, and cram them together on one chip. Again, smaller means slower heat dissipation. A single (larger) die should do better.
  3. The XP-G3 is known to be capable of nearly 2000lm when pushed according to djozz testing. Again, this is with a smaller die and a smaller package size than XM-L2.
  4. We all know that Cree emitters are highly under-rated. I think an official rating of nearly 2000lm for a XM-L3 is clearly possible with the technology Cree holds.

And following that, the reason for it being called XHP is obviously to distinguish the 4-die arrangement. It especially makes sense with the 2S2P arrangement of the 6V version.

I also don't necessarily see the unavailability of an improved 5050 1S emitter completely bad. It's time to bake a batch of efficient boost drivers. While not as simple, this easily means fully regulated emitters, 100% PWM free modes, and full battery voltage range utilization.

Also, time to cook low RDS(ON) high current FET switches. Plenty of small & efficient MOSFETs which can be paralleled for reliable & comfortable handling of 10+A tail currents at the press of a tiny switch.

Cheers ^:)

What the problem of boost driver are the coil and efficiency
More power, bigger coil

Not only that, but boost drivers put a big strain on the batteries. As the voltage difference (input => output) goes up, so does the current requirement from the cell. We like to overdrive our emitters, but to even reach rated levels with a boost driver would already be very taxing to the cell.

We’re gonna need good buck drivers for our modern “3 Volts” LEDs.
Same problem with buck drivers though: The coil needs to be beefy for high currents, driver needs to be efficient.

Linear constant current drivers have to burn off more excess voltage with those low Vf LEDs
LiFePo is getting interesting, with their more constant voltage of around 3.2 Volts. Still a bit low for the latest LEDs though…

But, meanwhile the modern LEDs are getting more efficient, so you would need less power for the same output.

True, but there are good high drain cells from reputable manufacturers available, even with 3000 mAh or more.

Well, you’re right about that, but high capacity and high drain are both more expensive than laptop pulls. :money_mouth_face: Seriously though, to need a high-drain cell just to turn your light on, and still not be able to push the LED very hard, just isn’t appealing to me. Maybe when those 21700’s that HKJ tested a while back become commonplace, and good high-output boost drivers are available to make use of them, then I’ll relent. What I think would really be great is for the Vf of the LEDs to continue dropping lower so that either LiFePO4 cells or buck drivers with Li-Ion make sense for single-cell lights, as you mentioned above.

  1. Any load from driver with regulated output is more pleasurable to cell than direct drive.
  2. Most configs can burn you hands with several minutes (even single 14500 cell in Cu X5), really usefull long-time mode wont overload existing cells.
  3. Manufacturing industry cells with high rates raises, while laptops are flowing to the li-pol batteries. $10 for four HE4 cells - not sure that you can buy new laptop cells cheaper.

maybe this is a stupid question, but why we need buck driver for 3V led, when you can simply PWM the voltage
if you want to regulate the output, then add the Rsen, a filter, and something like that, there will be no coil