For reference I tested the recent XP-G2 S3 3A build with a 6A ebay forward clicky. I’m just not getting the numbers he is.
It’s making me want to send him one of my tested lights and see what his meter measures it at.
I have read many times here at BLF, that warmer emitters generally in the same flux bin most often loses to colder ones.
It can be from the higher phosphor thickness or the not pro level light meters affinity to read colder light higher, or aspheric lights that often seem to get higher readings from colder leds.
Not a 100kcd difference though. I have an S2 2B running at 5 amps in a 1405 that only measures a tad over 400kcd on my meter. That’s the same tint, 2 bins down, at almost 30% more power. It shouldn’t be 55kcd lower than his S4 2B.
As i calculate it thats almost exactly the expected increase of 14% (400kcd +14%=458kcd)from 2 flux bins.
I know you said its at 30% more Amp but, the higher flux bins always seem to need less power/watt to get to max output.
If you check his XP-L V6 (that i believe is a V5) and compare that to the XM-L2 U4, that is also 2 flux bins difference (if i am right about the V5 part) and show a similar lower Amp because of higher Vf but a higher output characteristic.
EDIT, The difference of the XP-L to the XM-L2 U4 is almost exactly 3 full flux bins or 21, that would fit if the XP-L is in the lower scale of its bin & the XM-L2 is in the higher scale of the allowed 7 variation on flux bin steps.
But as you say the only way to be sure is to use the same meter, or that you send him a light to test 
The bin by itself would only explain it if they were running at similar power levels as I understand it.
I’m really starting to suspect that I have a meter that reads low, which is a much better situation than my lights all underperforming in a very consistent manner.
I am not sure it works like that, how would you explain the difference i bring up in my EDIT at the previous post then?
I mean the good old XM-L could sometimes pull 8-9A, but that doesn’t mean it would output what a XM-L2 U4 at 4.73A can, the higher flux bins with optimised for efficiency emitters, always seem to pull less amp but output more when running fully DD. They are just converting less of the energy to heat, but unfortunately they also seem to get higher Vf in the process, so they get harder to drive with our battery’s sloping discharge curves at high output for long, in single battery DD driven lights at least.
I guess you can draw some conclusions when you get your XP-G2 S4 from IOS and compare
What meter do use? Maybe its know to read low.
n10sivern and I have both built XP-L V6 3C versions of the 1504, and his measured much higher than mine there too. I’ll know more once I put one together with an S4 2B.
I’m using a Ceto CT1330B meter.
My mods are simple. Hot dedomed emitters with hot air reflow gun. Melt the solder on the pcb, continue with the heat and count to 20-30, then wait on solder to harden and dedome with a razor. I’m using a DD driver with 20ga wires. Artic Silver paste. Pot the pill with duct seal. 20ga switch spring bypass. High drain batteries.
Ok thanks :), good to know when i estimate where i could end up with similar emitter options.
“Pot the pill with duct seal”, is something i don’t know. How does it work?
I guess its for improved heatsinking, a pic would be interesting 
When you “20ga switch spring bypass” do you solder it to the bottom of the brass post or drill through it & solder it to the top?
I did that on mine because i suspect that long brass post, could be limiting current, of course not more that a % or 2 at most
But i know Dale have pushed XP-L’s up to around 7A (EDIT forgot to add, at start here sorry very different to at 30sec) & it seems like with these new high Vf emitters, every little bit of less voltage lost in the circuit, have a big additive effect in the end of max output & max amp.
I just noticed that your second XM-L2 U4 is about a full flux bin lower (9) in result than the first, that would almost fit in the allowed 7 flux bin variation, but the first one is even higher than that, so that one could be a flux bin lottery winner
I do both. Just depends. These have a small piece of wire soldered at the bottom of the spring and then soldered at the top of the spring. I then solder the brass button to the spring. I have soldered the wire directly through the board the switch is on. I’ve tried 18ga wire and saw no difference so I use 20ga as it’s easier to manipulate and fit in the spring.
As for the differences in the XM-L2 U4 emitters, you have to expect some differences in emitters after dedoming. It may not be completely obvious but there could have been slight damage to the phosphor or whatever to cause differences. Or it could just be the LED lotto, win some lose some.
2 seemingly almost identical looking 50mm aspherics?
one could me 10 or 15mm different FL, and while they will LOOK identical? they should perform differently enough to notice…
I still can not believe how bad results you got on small aspherics (worth to do test with another 1503? ). They are so bad that I would take a risk of finding replacement lenses all over the world if it will happen to me...
Cajampa you really must get yourself lightmeter man :) Although you did very good and tidy mod I think that 22awg wire is more than enough to bypass switch spring, and you don't need to bypass anything on driver if you use carobronze spring from FT. But who knows? Maybe I am wrong. I would like to see your opinion when you get lightmeter(even budget one). Because it will tell you if you are getting something or not with your mods.
I tried before with thicker wires(but of course I will admit I did not tried with currents above 4A in single cell light) and there is absolutely no difference. But problem I did encountered while bypassing with thicker spring that flexibility of main switch spring decreased(no flexibility at all) and it unnecessarily forced bottom driver spring to bend to its maximum or I used to have small dents in protected 18650 batteries which is not good.
When upper and bottom spring flex properly they works as one and handles any kind of recoil without damaging battery.
There are so many variables. A Cree bin is a few percent wide, but Cree guarantees their own measurements only within 7%, so that is on top of the bin-variation. Then the variation introduced by dedoming (which I have never measured separately but suspect is less than what is the general opinion) and quality of the reflow. Then the variation between luxmeters (not to be underestimated!), varying levels of clumpsiness of the measurers causing measuring errors, lens lottery.
Of course all variations will not be the maximum that is stated (manufacturers are careful with their claims, measurers may be have a well thought out and consistent method), but still I sometimes wonder how it can be that different people get comparable results at all.
10% variation is a job well done when measuring light!
I will someday soon, but those light meters KKW & n10sivern have cost like $30-$40, and that is about what i payed for the 1504 + XP-L on Noctigon i used.
I haven’t done enough research to know if there is a cheapie that is good enough to invest in yet.
You are right that there is less to gain at lower amp & non if you use a linear driver at a fixed amp.
But then again at higher amp on these high efficiency high Vf emitters, every little voltage gained & not lost in the circuit, from the battery to the emitter improves the output.
And i don’t agree that a carobronze spring is as good as a wire bypass, but a beryllium copper spring is almost there. To bad we only have that tiny one from IOS to chose from, i hope a full selection of copper springs in all sizes comes soon.
You are right that you can lose some spring travel but i think its worth it, for a for a max output build at least.
I used a carobronze spring, with a wire bypass, and the wire has free travel into the driver area in the pill for the led/bat + wire, but this only work when you don’t use a driver.
Are you building another one here? Because I don't see solder on a pill and driver.
And I also see that you put spacer for light to not go in over focus mode right?
You have excellent soldering skills and you are probably electrician or electrical engineer?
But anyway I would like to see driver soldered for that brass pill.
18awg wire in both spring is thick but it is probably necessary for high currents as you told.
What is expected runtime with your driver at max lux performance and lets say LG18650D1(3000mah) or similar batteries?
Edit:
I think that de doming process difference is another important thing that experienced members with right equipment could measure among same batch of same emitters bin.
We have 3 main: Gasoline, Acetone, Hot dedome.
Some guys says that gas de doming leeds to greenish tint while acetone and hot de dome not. So if tint remains cooler with other types of de deoming and since aspherics likes cool white tints I would not be surprised of performance difference between them.
I always reflow with slightly pressure to led dome.
I like hot dedome because it takes a whole 2 minutes. I’m impatient. I am also not a big fan of warm tints in throwers, just flooders. I said I was going to do a test with all 3, but I’m running low on emitters I’m willing to sacrifice. I will be making a mouser order next week, maybe I can find some cheaper BIN’s to test.
You will become a legend as Djozz if you continue this way... You guys are precious for me like that ring in the LOTR :)
Nah, No legend for me. Hell I’m still new to all of this.
Each man that sacrifices his time and money for testing, or gathering interesting lights for all of us, or any other useful flashlights theme stuff should be considered legend...
Without you guys and plenty of other guys that could be mentioned here there would not be BLF, and lurkers like myself would never learn anything.
About 30 most experienced people here made the whole world stuck to the screens of BLF and to join conversation. Experimenting and Sharing knowledge and your own practice is something that cant be replaceable by book and generic manufacturer provided data.
[Quote=luminarium iaculator]Are you building another one here? Because I don't see solder on a pill and driver.
And I also see that you put spacer for light to not go in over focus mode right?
You have excellent soldering skills and you are probably electrician or electrical engineer?
But anyway I would like to see driver soldered for that brass pill.
18awg wire in both spring is thick but it is probably necessary for high currents as you told.
What is expected runtime with your driver at max lux performance and lets say LG18650D1(3000mah) or similar batteries?
[/quote]
No this is the same one i built before.
The reason you don't see solder is the "driver" is just pressfitted in. But it isn't an actual driver it is just an empty 17mm spare ospark board i had.
Check out the led side
http://i.imgur.com/Mbecfni.jpg
If you look at the LED - you see that it is soldered to the brass pill, there is no need to solder the bat - on the driver board to the pill. The electrons already goes through the bat - on the battery through the switch in to the tube & in to the brass pill that is screwed into the tube.
I don't have to use a led - wire or ground the driver board, because i have soldered the led - to the copper mcpcb & the brass pill ;) for maximum flow of current :D
Yes i put some tape to adjust the focus, my 1504 with the XP-L on sinkpad, is a quarter of a turn over focused when fully zoomed out in spot mode. It isn't perfect yet with those spacers but its a bit better. I will change the led to a XP-G2 S4 soon anyway.
I actually am a trained electrician, never worked as that though.
And slightly embarrassing to admit to 
because i am so bad at much of the basic electrical math, i was supposed to pick when i went to school for it. But at least i am certified, so i can rewire stuff & do basic installation work legally & that is always nice, when you want to fix stuff around the house.
If you thought that spring looked nice soldered, it is a trick that Dale shared, i coat the bottom of the spring in solder paste, and just reflow the spring with a heatgun ;) works great :)
I don't know how long a LG18650D1 would last, because this is fully direct drive from the battery without any protection, it is a bit dicey to run it to a fully drained battery.
The only protection is me making sure i don't let the light dim to much before i pop in a new battery, but as long as you always check the battery with a DMM after use, you quickly get a feel for how far you can drain them safely ;) You could use a protected battery, but the protection circuit eats up a bit of the max current & voltage, but at least it would be safer.
I got a "Panasonic NCR18650PF 2900mha":http://lygte-info.dk/review/batteries2012/Orbtronic%2018650PD%202900mAh%20%28Black%29%20UK.html , i could test how long it takes to get down to about 3.2v or ~80% drained, i would guess about 15-20min or there about.