this is probably posted somewhere, but I was wondering how many 7135 chips can you use on a 18650 light with either an xm-l2 u4 or xp-l hi v2? 11 ? 12? for max output. thanks
Michael
each 7135 adds either 350 or 380ma. You can stack/add as many as you want, but eventually your battery won’t be able to supply that much power.
I know as much as you said, I am wondering what the actual limit is? is 11 too much 12, 13 where is the point where the best battery is not going to help it get any brighter? thanks
It really depends on the emitter. You used to be able to easily get 6+ amps out of a single XM-L/XM-L2(16+chips). Now, on a single cell most cree emitters will top out around 4.5-5amps (12-13 chips). On a triple or quad you can do a lot more than that.
Alright, scratch that.
I wouldn’t do that……
It wouldn’t tell you anything anyways. You need an LED in the circuit to get useful information.
I see you posted while I was writing, and that looked like good info. Although I think it’s the only way to discover the battery’s peak output. I guess you would want a voltage reading at the same time though.
even the cheapest 18650 will do dangerously high amps when shorted like that, but we want to know how high they will go with an led attached. It’s still a simple test if you want to try it, just keep the LED cool somehow.
thank you, that was what I was looking for. how many would you guess on a triple xpl? and how many would be worth it on a convoy m1? what about xm-l2 u4 in an convoy m1? thanks again,
Alright, I edited the post.
6A is tops for the led before the bond wires blow on the newer XML-2 LEDs. With spring bypasses and a top 4.2V cell this is possible but only briefly. Keep in mind the more current you pull from the battery the faster the voltage will drop and fall out of regulation(start dimming). This post by Djozz should clue you in to what you’re looking for. 12-13 chips with good heat sinking and low resistance mods(spring bypasses etc) is doable.
great! thank you for your response.
One reason the triples are popular is the better efficiency and lower Vf of LEDs in parallel so you get more lumens per watt and longer runtime in regulation at a given current than a single for the cost of a floodier beam.
I can’t remember where he posted it, but Dale (DB Custom) has said that he has a triple that reached 21 amps from a single cell. It was a high-amp cell. I think an Efest Purple.
Edit: Oh yeah, it was a triple XP-L, I believe. You probably wouldn’t want to try 21 amps to any XM-L2’s, especially the newer ones. They will definitely blow. 
If I was building a driver, I would use a FET for the high modes, and some 7135’s for the low modes. That gives direct-drive high like seems to be the goal here, and it gives constant-current for the low modes, which is best too. You get the best of both worlds, and probably save space too. You can thank me now 
does the fet do the high mode only while the 7135 does all the modes below that?
Nope, it was a triple Nichia 219c. The low Vf is why it was able to pull so much current.
The FET should be PWM’d to cover the range of 100% down to where ever the 7135s can pick up. For example, at 90% your PWM’d FET is still very efficient, and trying to get enough 7135’s to achieve 90% is going to be a waste of space too. I’m thinking a FET along with as many 7135’s that can be comfortably added. Then where-ever the 7135’s top out (75% for example) the FET takes over for settings above that.
great! that means I could have the fet do hi/turbo and then my next mode up would be 50% then 30% then 5% then moonlight. all of which could be run using the 7135’s. is that right?
Not really. Speaking in percentages of power, a single 7135 will not get you anywhere close to 50. More like 5-8 of the total Fet power. So moon and 5% could be on the 7135, the rest on the FET or a combination of both (which is essentially how the BLF A6 works). To get the rest, you need more 7135’s.
(that’s why I made my DoubleDown driver)