If only I could predict the lotto numbers now!
The individual emitters sold out weeks ago. There are still some on stars but I don’t know if they are old or new. I plan on ordering a few to check but need to wait until I get paid in a couple of weeks.
I always claimed that same kcd values of reflector vs aspheric are not same… With aspheric you’ll see further in a night.
I have 340 kcd reflector and visually it does not out throws even mine 250 kcd UF 1503. Both measured on same light meter at same 7m distance.
So is this perception(the way we percept things) or beam spreading at a distance I can’t tell exactly.
I got my 8 good, old and precious XP-G2 S4 2B and messed 2 of them by by stupid recklessness so Smeagol is very angry now 
Yeah thats true m8, yet refclector based torches have their atvantages, even if the target moves from the hotspot you still can spot it cause of the ” oreol, a.k.a halo” lighting caused by the reflector- and its very efective in open filed type hunting.
And ofc thet halo lighting “nukes” your scope eye if there are leafs, grass and other reflecting back obsicles…
By its another topic, and its hella long
Cutter still as S4 2C- its also a great emitter, same performance, same amps
I recently bought these S3 2B fom TaoBao. And have one of them changed into my Courui (DD) and found that it is actually brighter than the initial 2016’s S4 2B. However I still not sure whether these are old version.
They were produced in Dec 2014. Anyone know when CREE changed their production with new batches?
With a date of Dec 2014 they will be the old version.
Noted. Thanks 
One macro shot tells more than 1000 of words. Very easy to detect them…
Can you report some kcd values of your Courui with S3 2B? Yes it probably is brighter than new S4 2B but even old R5 2B is brighter than it.
Old and better version of S4 2B is different animal.
I will quote myself:
“Very simple to detect them by dot method.
Old has around 40 visible dots on emitter while New has more than double number of mentioned dots and they are very hard detectable.
”
Yes, this is the old batch.
From the left to right: XP-G3 S5 2A (May 2016), XP-G2 S4 2B (Mar 2016) and XP-G2 S3 2B (Dec 2014).
(sorry for dirty LED doom surface as they are so small to handle with).
Very nice photo.
Everyone can see difference.
I just got a new style XP-G2 S4 2B from Richard. Tested it in an X6 with a fresh 30Q and got 4.05A and 52,700cd with dome on. Then I gas dedomed it and got 4.2A with a fresh 30Q and 98,800cd. The focus may not have been perfect when dedomed. With dome on the yellow phosphor filled the reflector at arm’s length. When dedomed the very outer part of the reflector was not filled with yellow. I think I might need to file the spacer down a bit, but I didn’t fiddle much with it. The tint seemed similar to the old S4 2B.
This performance is pretty close to the old style. Comparing to other peoples results, it seems like there is some variation in Vf for these. Maybe there is not much reason to be sad after all.
Also, I tested an XP-G3 S5 3A I got from Richard in the X6. With a fresh 30Q, I got 6.7A and 58,600cd, with dome.
Well when you have “only” 100 kcd performing light you will certainly not feel those 30 or more % performance drop. That should be 30kcd max performance difference.
But deduct 30% from 250 kcd light and that should be 75 kcd difference.
In a 1504 (around 500 kcd rated with old xp-g2 s4 2b ) that will be 150kcd
etc…
So IMHO there is really a reason to be sad because instead of better performing emitters in 2016 we got… What we got? Nothing… Even with XP-G3…
So I was going through some old flashlight parts, and I found 2 (apparently) coveted XP-G2 S4 2B’s on 20mm Noctigon’s that are the “old” style. 
Looks like I’ll have to hang onto these puppies until I find the right build for them.
I’ve been trying to keep up with this thread, but have slacked off for the past few weeks. Is it correct to assume that the new XP-G3’s do not out throw the “old” XM-G2 S4’s when dedomed? I was considering trying my hand at dedoming a G3, but some people say they do, others say they don’t.
Stick with proven thing for your thrower
Read this: Cree XP-G3, testing a S5 3A emitter - #79 by luminarium_iaculator
Not at all, even completery clean sliced or dedomed perfectly the new G3 has lesser performance that the old G2s, like….60% lower or more
I dedomed like 20 of the new G3s, have several nearly perfect ones, the best one gave 213cds in B158, the new S4 2B ( at 4.8 amps) 260k and an old style S4 2B ( at 4.6ams) 330k cds
Ofc G3s could change in the future, there might be newer revisions with a new core dimmentions
This confirms one of mine theory I talked into XP-G3 thread and that was about neatly sliced XP-G3 vs de domed G3 emitter where I claimed that there should not be great difference between this methods.
So sliced(neatly sliced as low sliced as possible) vs de domed = very low advantage onto pure de domed one.
Indeed, slicing is needed especialy in b158 look loke hosts where ppl usualy swap the emitter/driver pill on demand…with an IR or red/green for instance
Ofc you can always use a HI in those……IF they dont die that fast as they do :)…especialy those tints with lower K rate( 4500 and under.
By default they are OK though…yet after 5amps the sillicone layer near the bond wires starts burning or they just suddenly die
Well since S4 2C is only available in “new improved version” I played a little with remaining old XP-G2 S2 1A emitters…
And guess what?
One of them hit 5.8A!(checked several times with pause on unmistakable Uni-t clamp meter) Angry blue appeared on half of emitter (I did not fried it… just a second or two), and that is something I never saw with XP-G2 emitters… Samsung 30Q used.
Then I had to change Djozz FET driver bypassed spring for original UF 1503 high resistance spring (reading 5.4A after), and unsolder tail switch bypass wire, and angry blue disappeared for 200 kcd results in UF1503( not bad but nothing special).
First time I had to enhance resistance on Djozz FET drivers… I don’t know why but it was funny to me :laughing: Am I nerd?
What I know now that Djozz FET driver takes out maximum out of LED and battery if done properly (filled vias on AK-47 + bypass from that vias to top of the chosen driver spring(FT carobronze) gives direct + contact with battery and led)
has anyone tested those : http://intl-outdoor.com/noctigon-xp16-v2-mcpcb-cree-xpg2-s3-3d-led-p-846.html if they are old style ?
If the reel information picture is correct and updated: the package date of the S3 3D is april2015, while the S4 2B (that is currently new style at intl-outdoor) has an earlier packaging date of februari 2015.
Djozz,
Hope you’ll test and find even better emitter than old style G2 S4 2B