how far can i go with stock LED on ALU ?
I tortured the stock board+led in place (so with the original thick layer of white stuff thermal glue) with a NANJG92 DD driver for short bursts at a time, I did not measure the current but later with a XM-L2 on a Noctigon it was 5.5A. When I dedomed the stock led it showed burnt silicon around the bond wires, that is when I decided not to use it. It did light up alright with my led-tester (50mA).
Tonight I tried a dedomed XP-G2 R5 1A (from Fasttech) in my UF-1405 shorty with NANJG92 DD driver. The tail current was lower than with the XM-L2, even lower than expected, 4.6A at 30 seconds after switch-on. The throw disappointed: I expected over 500kcd, but got 407 kcd at 30 seconds after switch-on. I would expect that also from a good dedomed XM-L2 ! 
I'm not sure how this happened, apart from me plainly doing the wrong calculations, several reasons come to mind:
-it was a cheap R5 led emitter from Fasttech with perhaps by random a relatively high Vf. The current is quite low here, 4.6A compared to 5.4A in a similar mod with the same cheap led of the UF-1406 (5.4A current alone would still not get me to 500 kcd btw)
-the careless reflow I did this afternoon was too hot and damaged the output of the led a bit.
-there's a bit of current leaking around the led via the (anodised, but touching the solder blobs) led tightening ring.
Well, that was the experiment, I'm not going to finetune the XP-G2 in this mod, the final led will be a dedomed proper XM-L2 if I can find a good one (I'm in no hurry)
Why you tried this old sucker? :)
Use XP-G2 S2 1A from Vestureofblood and expect 15-20% improvement, and why drive it more than 3.6A?
I think it starts to loose performance over that point if it is just on noctigon + brass. So with S2 if 20% that will be around 500kcd but even 400kcd is killing throw.
407kcd is very good already, don’t forget you are using 66mm aspherical lens, it’s not small but it has its limitation too. 
XP-G2 Vf is higher than XM-L2, thus with only one battery used and direct drive the current draw will somehow limited at some point due to the voltage sag of your battery. High current buck driver with two (or more) batteries can overcome this issue because we will have a lot of input voltage. Ahhh well… until we have a good 17mm buck driver we will talk about this again, for now a single 18650 and 17mm Nanjg driver will do the job pretty well. 
I used a 17mm contact board and 18awg wires to the LED, so it is now a one mode direct-drive light.
Current = 5.5A at tail (single 18650)
Throw = 290kcd (only?!) converted from 4.5 meters.
Why my lux number is so much lower than djozz’s with similar setup? I am suspecting my XM-L2 U2 is a bit beaten up, and also is it because my measuring distance is too short?
Direct drive wiring the battery straight to the led is always a great idea :-)
Measuring from 4.5m compared to 5m should not make a measurable difference if calculated back to 1 meter: different than a reflector light an aspheric light does not need a considerable distance to collimate, as long as it is focused correctly.
I don't know why your number differs. There's more variation in luxmeters than we like to know, but 25% off is quite a lot. Further, the correct way to measure is moving around the meter a bit inside the hotspot and look for the highest reading, but also that should only vary a few %.
I am guessing it is the LED. It has been reflowed for several times and used in my T08 before, somehow it is damaged?
I can’t wait for my XP-G2 S2 to arrive and put it into my light.
Another factor that may be of influence: Luminarium ejaculator has reported that different copies of the Uniquefire UF-T20 vary considarably in the quality of the lens, with a measurable difference in throw capability. The UF-1405 has a glass lens as well, so perhaps they also have this varying quality
:) Luminarium iaculator = thrower flashlight on latin :)
And not only UF T20. Performance vary even among premium aspherics like Dereeligh, Crelant and Tiablo. Deree NMV2 and Tiablo are same size lenses with different focal lenght.
2 lottery things happens when you buying aspherics lights:
1. Aspherical lenses lottery +- up to 15%
2. Emitter lotter +- up to 15%
So if you can get at least +10% with selection of lenses + 10% on hand selected emitter that is killer combination.
Djozz you are BLF if you ask me. So much work with pictures is facinating.
Could you test performance of emitters on noctigon + regular brass pills? I really think XP-G2 starts loosing performance if driven over 3,5A on generic brass pill. In fact I don't see sense driving it over 3A for single nanjg.
I just realized one thing that I didn’t use copper-braid on my switch’s spring.
So I went on and used a pair of tweezer to short out the tail (reduce resistance) and this time i got 310kcd. I think the tweezer did not give enough contact too, maybe with copper-braid on it I can get even higher reading if not close to djozz’s one.
Try silicone flexible 22awg wire(noodle) instead of copper braid.
Much easier to solder than copper braid(at least to me) and spring remains relativley flexible.
I used wires before I bought copper braid. I think in terms of flexibility they are about the same, only thing is with the copper braid you can’t let the solder flowing through the whole braid otherwise the whole copper braid will become very stiff.
thats what happens to me when braiding… 
either it won´t stick to the spring or it turns into a massive monument…
My way of doing is nothing special at all, put some flux just on the tip of the braid because this is where you gonna attach it on the spring, when you put your soldering iron on it remember to release your iron quick enough to prevent excessive solder flowing into the braid.
I've done a ton of mods with braid - many failed over time. Now, I use nothing but wire - either 22 AWG if I can, or 24 AWG. I haven't seen one wired spring fail yet. The wire insulation acts as a barrier to prevent the flux creep down the wire, and the insulation also keeps the wire from breaking with heavy use of flexing. If you use lots of solder at the top of the wire (top of spring) like I do, even if the solder comes off the spring, chances are the wire will still be making contact with the battey. The braid failure I've had is usually from it breaking in the middle, so it does seem to break the braid connection, even on compression - it seems like.
Hi guys, I’m a bit new to this torch thing, however I’ve lot’s of experience in electronics field.
I read this thred from top to bottom and I’m about to buy this 1405 torch.
so I’m settled at the torch and I would like to go with the XP-g2 s2 1a (noctigon star) @ 3.6A, so my guess I’ve to try if there’s a good driver to handle 2*26650 right? any reccomendation here?
about these 26650 batteries what do you recommend? I don’t have any so I want to buy some good ones…
best regards
Simply put, there are not buck drivers that will fit (17mm) and will deliver the power you’re looking for (+2.4a) that are available without making modifications (to my knowledge) . Many of the folks running high amperage drivers in this thread are doing so with 1 battery, either through dummy cells or by cutting the light down.
well I guess I’ll have to wait for the torch and than see what can I do on it.
I’ve a fairly large workshop with Drill press, band saw, lathe, CNC milling machine and lot’s of other stuff so even if I need to make a new pill or something I can do it.
meanwhile I’ll be reading lot’s of threads here on this forum 
what about the batteries what should i get?
Try to make us your version of wavien collar for this light.
Borg from Romania managed to make it but he affraids to sell it. He thinks that wavien will persue him if he sell me one over BLF pm system :)
I cant talk him to sell me few of those. When you will stuck into this hobby and since you have so much machines please try to make something similar as Borg at this thread post 51 https://budgetlightforum.com/t/-/17005?page=1