If anybody can help out with any old xp-g2 s4’s please let me know. Im looking for any qty available…
Mee too 
LI has got them all now 
I’m about to remove the S42B from my meteor and replace them with XP-G3s, my meteor is from the first batch they should be the good stuff right?
I don’t commercially mod flashlights so perhaps I can sell a couple!
Edit: I must first wait for a CW XP-G3 and know for sure the output under a carclo is better before swapping them… Will let you guys know.
If you want contact me please Will .
Thank you
Yesterday I tried few old(good old) XP-G2 S2 1A dd. They all pulled around 4.8-5AA from fresh 30Q but they were unable to bring performance of older brother S4 2B at 4.4-4.5A (around 200 kcd in a light where S4 2B brings 250kcd)
Only good(or maybe not good to someone) with S2 1A is that it still has clearer upper neutral white tint (more white in a tint than with S42B, S2 1A is without green pigment at the edges).
So I am really looking forward for my incoming S4 two bees 
Not every 2b emitter could do that in a 1s setup and a fet driver,
In the matter of fact they tent at 4-1-4.6,
In about 15 days i will test the old vs new S2 1A , and report back here.
Hhmm, thought I had only new S4 2B's, but seem to get 510 kcd in a single cell UF-1504. Got 4.25A at the tail on a EFEST 4200. The batch of 5 XP-G2's S4 2B ordered/came in Feb 2016 from IOS.
used the "hot gas" method, on a gas stove of course. Very safe, cooking gas in a pot of water over gas flames. 
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.
old S2 1A can hardly reach(it can but only if you are lucky with emitter pick) performance of old S4 2B.
And old S4 2B is better for FET drivers cause it sags less current(only 4.4A or less draw) while S2 1A will eat about 5A of current with lower lux performance than old s4 2b (but nicer tint)
So if you are going to compare old S2 1A vs new S2 1A that will be interesting….
Ok, just checked - it's 40 dots - definite, I counted them, or darn close. 5 rows of 4, 4 rows of 5. Order date of Jan 20th, 2016, shipped Jan 24th, rcv'd in Feb 2016.
Does this make sense?
Yes
So what Hank was shipping in January was good? I ordered mine bare.
My new ones were ordered at 13 February 2016.
Hhmm, guess I could have gotten lucky. Maybe the bare ones were remains from an old batch, or the new batch came in end of Jan, beginning of Feb. Sometimes he says the LED's are from a new batch or not, I've noticed. He's also very helpful when asked. I've asked couple of times for LED's bare that were not listed, and always were able to get them from him.
I should re-check tail amps then with my clamp meter - might be a little higher than 4.25A. I'm using 20 AWG wires and a FET+1 driver with the SIR800DP, springs bypassed with 22 AWG wires.
Okay… Good info on the number of dots. I guess I need to get my microscope out and check the few XP-G2s I have in my kit. I have a number of older emitters, but mine are all likely 1A tint.
2x magnifier is good enough for this.
I realised which of them where old or new , according to the size of the die (with help from li).
It’s pretty obvious if you learn what you are looking for.
I used a cheapo 10X pocket magnifier, along with my 2.00 drug store glasses to examine the dedomed LED, and that did the trick for these 58 year old eyes. I find magnification combined with a good light source really, really helps. My hands are still pretty steady. I work with a guy in hist late 60's and be does some super small incredible prototyping work still, but same thing - having the right tools is huge.