XP-L2 V5 Output & Death test by Texas_Ace Over 2200 lumens! Still worked after 15 Amps!

It is possible, it is just not spelled out in the data sheet. Very possible it is just some guy saying that looks like a 2D and selling it as such as well.

Thanks for the testing TA!

Will people hate on this emitter like they did with the XPG3?

What your average output numbers for a triple S2+? This must be coming close to those numbers?

I have a feeling people will hate on the tint shift like the XP-G3 since the tint shift is basically the same for both of them.

Although the extreme output these have should off-set that in many peoples mind. I personally never really minded the tint shift in the real world myself, you don’t really notice it unless white wall hunting.

Far as output goes, it doesn’t quite match a triple setup but it is a lot more usable IMO since it is not pure flood and actually has a beam and hot spot. With my XP-G3 90+ cri mixed tint triple I am getting around 2500 lumens without a bypassed rear spring (been too busy). I assume it will be around 2800-3000 lumens with it bypassed. I will actually be selling that light before long.

The XP-L2 looks brighter though since the hot spot is more intense and that is what the eye notices.

Overall the XP-L2 is great for output, the tint shift is something it looks like we are just going to have to deal with as we move into the next generation of LED’s. Even the 219c’s I got recently have a slight tint shift on the edge of the beam (although much much less noticeable then cree).

I'm thinking the best use of these would be in OP reflectors or triple/quads, either TIR optics or SRK's, etc. I'm thinking roughly it could add 25%+ output to the BLF Q8, maybe 50%+ output to triples and quads like with the X6's many of us built, besides huge output gains in standard 20 mm triple lights.

Of course the output bump is mainly attributed to higher amps, which means higher heat.

What I'm not understanding is the higher output at fixed amps of 350 mA and 2.45A. If it's bin 5, wouldn't that match output at fixed amps of a XPL V5? But you are clearly seeing higher output lumens than any XPL V5 LED, so, why would the CREE specs indicate otherwise?

The best # I've measured on one 350 7135 is about 180 lumens, but that was in a triple or quad.

Tint shift is a killer for me. If they make an XP-L2 HI somehow I’ll probably jump on board, but I’ll give up quite a few lumens to not be disgusted by the beam when I turn it on.

Also, I think Cree themselves do label things with specific bins like “3D”, but it comes after awhile for a new emitter. It seems they try to get the first few batches out quickly, then do more specific binning later.

This is all based on my fairly vivd memory of seeing Cree labels with specific tint bins on them in the past.

I think the largest reason for the numbers was actually not that the numbers were wrong but the emitters were wrong. Up till recently I have been collecting parts but kept waiting for the TA drivers to be finished, then TK to update the firmware and finally for the funds to actually build them.

A lot of the lights I have built so far had LED’s from “sketchy” suppliers, My guess is that the bins were not correct in them (big part of why I didn’t want to test them and base results off those numbers).

I have been building a LOT of S2+’s this week with various setups, here is a random sampling of numbers from them to give an idea. These LED’s I am sure are genuine bins but they are committed to gift lights so I didn’t want to risk messing them up.

Oh, the numbers from my test actually line up perfectly with the Cree data sheet when I compared them via http://pct.cree.com/dt/index.html

219C 4000k 90+ CRI D240 from group buy

350ma = 140 lm
2.45 = 740 lm

XP-L HI V2 3D (the best throwing S2+ I have with a TIR)from IOS

350ma = 170 lm
2.45 = 1000 lm

XP-G3 5700K 90+ CRI from Arrow, not highest bin

350ma = 193 lm
2.45 = 930 lm

XP-L2 5000k V5 from mouser

350ma = 200 lm
2.45A = 1150 lm

So with known bin emitters the jump is not as much as I originally thought, It seems that fasttech and others can’t be trusted with their bins :person_facepalming:

I should have noted where the other LED’s came from.

The biggest change is what happens when you keep pressing the LED and the fact it can pull a lot more power from the cells, even cheap laptop cells can give nice output.

single cell buck drivers would also make a fair amount of sense with these LED’s.

I am debating slicing the dome on the one I have, the beam could be tighter for what I do with it and I am curious to see if it would improve the tint shift.

Although in the real world the tint shift is not all that bad, only in the house do I notice it.

I have some 4500k 90CRI G3’s that look terrible both indoors and out. The main beam is more like 5000k, and the spill closer to 3500k and a nasty green tint. The L2’s would have to be much much better for me to consider them in a reflector light.

I really hope this style of LED isn’t the new norm just for more output.

Sadly I think this is an effect of “flipping the die over”, aka not having bond wires. Since all the LED’s that have that also have tint shift, although nicha seems to have figured out how to minimize it.

I'd hope the nasty green tint is more a result of the true tint of the LED, hhmm, 4500K is maybe an E4 kit, so maybe it's a nasty 4C tint? We are rolling the dice with the tint on these E3, E4, E5 kit ratings.

I’d agree but the nasty green on my G3 was only the spill. Otherwise the center beam was mostly acceptable, though not actually 4500k. These are from Cutter, so no question of accurate bin reporting either.

My theory:
It is not the flipping of the die that does this, I think, but the type of phosfor, Cree used to have very fine-grained phosfor layers and was not very good in pleasant tints and better CRI, now their tints are improving and that coincides with rougher, more loose phosfor layers, that also are better in photon extraction at high angles (these angles in the old type of phosfor layer emitted much less light, photons were at the exit reflected back into the die, which was enhanced by dedoming the led) unfortunately that light at high angles has a higher proportion of longer wavelength photons and it creates the huge ugly yellow-ish corona around the hotspot (which is not spill light but light coming from the deepest part of the reflector).

Btw, TA, I stole your address from Clemence’s picture (a bit un-ethical, I know :innocent: ) and have sent you a small flashlight with very constant measured output (after 10 seconds basically constant, within 1%), so you can compare your calibration to my ‘djozz-lumen’. Not because the ‘djozz-lumen’ is neccessarily any better than any other people’s standard (I never had it officially measured, it is just my own baseline that I keep as constant as possible over time), and not that I wish you to use it (your standard may well be closer to reality than mine), but just to get an idea on how your emitter tests compare to mine in the past.
I hope it arrives at some point.

Great idea.

Thats a fantastic idea!

I was hoping at some point we could test the same emitter (perhaps the 219c’s we are getting from the same reel) but this would be even better. At the very least I could apply a multiplier to the emitter tests so that my test would be comparable to your back log of tests.

I see some PM’s I am guessing you are one of them, I will read them in a min.

The phosphor layer could be the cause as well, it is really hard to say since the phosphor change happened at the same time as the die flip.

Although it appears everyone is going to this new style, the nichia 319 looks interesting because it might have the high output without the tint shift.

I really have some hopes for the 319A, not that it will beat the XP-L(2) in output but that it will produce very pretty beams, also under Carclo optics.

I’m considering painting the deep part of the reflector matte white sometimes, so that it will mix with the spill.
Or just use a matte white spacer gasket with a high collar.
Not sure yet…

Anyone tried slicing the dome on a G3? Does it improve things at all?

Paining the lower part of the reflector is an interesting idea. I have a ton of spare S2+ Reflectors from swapping in TIR’s lenses, I might try that if I can come up with a paint to use and a way to apply it.

The light has a 4000K 80CRI (Samsung LH351B) led which strictly spoken just gives a correct comparison for that that specific spectrum (can be translated as: colour temp/CRI) and does not see possible differences in wavelength sensitivity for our set-ups (which is dependent on the quality of the reflecting inner surface of the measuring device and on how correct the wavelength sensitivity of the luxmeter follows the agreed Vlambda curve). In other words, a cool white led and a very warm white led may have a somewhat different multiplier between us than the 4000K 80CRI led that we calibrate at.