Thanks for your thoughts. When I get a chance, I’ll have to order a bunch of those emitters. Looks like it’s time to retrofit all the lights I’ve dedomed. Never liked the dedomed tints anyway, strikingly ugly.
well since the lumens per watt of XLP HI is basically the same of XPG2 as shown by your table, I’m assuming the HI are drawing more watts or there is no way they can put out significantly more lumens. am I wrong?
Yes it is the same for those exact bins, I mean XP-L HI V2, and XP-G2 S4, well the XP-L HI V2 still has a tiny bit more lumens per watt (3lm)
It's the same watt output (and it is capped, since it is regulated), so the same heat, the lumen output is just a another thing, an effect, how efficient is each particular LED model will decide the lumen output.
I believe the confusion comes from Dale, which for his XP-L HI U5 flashlight (yes U5 not V2) shows around 5500 more lumens than we expect, considering 10% optical losses ( that it should according to datasheets to measure 12000lumen OTF. I am sure explanation will reach the same level of wow-ness as the 12000lm.
Normally for the XP-L Hi U5 there should be a 7932 bare lumen output (not OTF) at 50C, so heat taken into account as the LEDs will not be as bright as 25C.
It may be voltage-limited instead of amp-limited. The XP-L HI has a lower Vf than XP-G2, so it would put out more amps at the same voltage.
indeed. that is why I’m expecting it’s drawing more watts, hence more heat.
lumens per watt are known and very similar to XP-G2, so it’s impossible that it’s putting out 12k lumens OTF drawing the same power of stock M43. whatever the explanation is (given Dale’s measurements are correct) it must involve a higher wattage.
You may have mentioned some of this before but:
1) Do you have any ‘throw’ measurements (for the XPL-HI mod) similar to the ‘peak beam intensity’ in cd used by the manufacturer? I was wondering if your XPL-HI had as much throw as the stock XPG-DD.
2) In brief, how did you perform the solder reflow to put the new XPL-HI on the MCPCB? Did you clean off the old solder with copper wick, etc. and where did you apply heat using the soldering iron to prevent heat damage to the LED, type solder used, etc.?
2) Do you have pictures of your spring/wire mod? Did you use 2mm wide copper braid (solder wick)?
Thank you for your input!
cncyana, pretty sure I’ve posted the throw… 51.5Kcd as I recall, something like that.
I have learned to slide an emitter off the mcpcb parallel with the pads to preserve the solder already present, I use a SS plate on a smooth surface stove top to re-flow then set it on a 3x3x3 6061 Aluminum cube to cool. I use Kester solder paste.
I don’t use braid for spring bypasses, prefer to use the same wire as I use in the lights for this. 20ga in this case, Turnigy Silicone wire. I’ve also learned to solder the wire at the base first, then put a slight coil to it before soldering to the top of the spring, this allows it to compress without bending or overly stressing the wire.
I think I replaced the 18ga wire with my own Turnigy 18 ga, might have used 16 ga, don’t recall exactly.
Edit: Wishing now that I’d taken current and voltage readings on the XP-G2 S4 2B’s that came with this light for a direct comparison. Bottom line is, the same equipment measured the light with the S4 2B’s at ~7690 lumens, the new HI’s do ~12,530 with everything else being equal.
you had the springs bypasses already when you measured the XP-G2 at 7690 right?
Are they? I remember djozz had issues with that.
Yep, both counts. Good battery and the springs were bypassed with the S4 2B’s.
And remember, my lumens and lux readings almost perfectly matched the ANSI specs from Olight on the M3XS-UT Javelot.
Sounds like you got this LED-replacement process down good! Just a few more questions:
1) I assume the original solder used for the LED/MCPCB is mixed with the new solder paste you use and this mixing doesn’t cause any problems?
2) Your method preserves the original solder - is this to simplify the process by avoiding the extra step of removing the original solder or does the new LED adhere better when you preserve the original solder or…?
3) I assume you solder all 12 LEDs at the same time when you heat the MCPCB copper plate?
4) Do you have to apply downward pressure on top of the LED so it doesn’t ‘float’ on top of the solder? I suppose ‘floating’ is only a problem if there is too much solder.
5) Has your process ever resulted in a burned LED?
6) What temp do you set your stove at?
7) I assume the solder paste you use is the softest/lowest temp solder?
8) When you remove the LEDs, do you heat the entire MCPCB copper plate and peel them all off one after another, then without cooling, repeat the same process in reverse to put the new LEDs on ?
9) Do you use a soldering iron to solder the bypass wire to the MCPCB copper plate toward the bottom of the spring? If so, can this same method be used to solder a new LED to the MCPCB copper plate rather than heat the whole plate up on a stove - maybe use a soldering iron to heat the MCPCB copper plate locally only - I assume reflow is best but cooling down that huge copper plate is a slow process while the LED is exposed to that heat during that slow process? Just wondering…
Once again, thank you for your input.
If that is the case than the output will all over the place, no one regulates LED based on voltage output instead of current.
Yes and no spring bypass will give you 5500 more lumens for the same output watts as before.
@Dale&others: how about building this cheapy sphere? Built in one spare evening, works as good as any sphere out there (better than most I dare say) and if you use a bigger styrofoam ball instead of the small one of my example, you can make a big entrance hole, and the range will be sufficient for any monsters you decide to build in the future.
well it wouldn’t be the same output watts as before, always for the same reason. springs bypasses increase the wattage by decreasing the voltage drop. however I obviously agree bypasses can’t provide more than a few lumens, I still see no explanation for that +5k though
I slide the emitters off in a re-flow action to preserve the solder already on the mcpcb. I don’t have experience removing solder, have no solder wick or vacuum or any other way to remove it and am inherently lazy. So I use what’s there, figuring it was good enough for the 1st one it’ll be good enough for the second as well. I PREFER to use a fresh star and a solder paste mask. But I don’t always get that option. I don’t like pressing down the emitter as it has often caused a short to do so. These small emitters aren’t heavy enough to push solder out if there’s too much there, so the right amount is much more critical. And no, I don’t think a soldering iron could be used to replace just one emitter on this large copper board. Hence, when one of the 12 wasn’t seated properly I pulled the board and re-flowed in the plate again. I use a setting of 3 to 3 1/2 on my stove top. I like to have the items being re-flowed in place when the plate reaches temperature, much more successful this way than having the plate sitting there full hot when starting out. In this way, only just enough heat is used, not too much, as it’s difficult to control these self adjusting burner elements.
I have modified easily 200+ lights (over a 100 of my own), maybe 400 emitters re-flowed in various styles (some lights have been rebuilt multiple times). I have yet to burn an emitter doing this. I’ve popped the masking off a Noction, browned the mask on a SinkPAD and never had an issue with the emitter itself. I’ve killed emitters de-doming, like most everyone else. I’ve burned up 8 in a row trying to modify a TK61, poof poof poof… too much spike when too close to the top end in current. Haven’t ever had one not light up after re-flow because it got too hot, but then, I always try to err on the cool side to prevent just such a thing.
Maybe the easiest way to prove or disprove my result here is for someone ELSE to put 12 of the HI emitters on this big mcpcb. Then we’ll all know what I already know. The output is staggering and the lightbox/meter are useless when out and about using the light. Some want to act like the numbers are what it’s all about, when in fact they are only a reference to what you can expect when in use. It’s really as simple as that. Datasheets are helpful, but again only guidelines with +/- tolerances easily responsible for big swings in output… especially in multiples of 12.
After buying 24 XP-L’s to test on this light (Yes, 24, remember that I sliced the domes off 12 to simulate the HI before it became available) it’s pretty frustrating to be so far out in expenses and be met with all the cynical attitudes. In the end though, I built it for me to use and play with, anyone’s disbelief is totally irrelevant to my enjoyment of clicking the button. Anyone wanting to know the truth can drop the coin on some emitters and find out for themselves.
Happy modding! 
You know how the saying goes, extraordinary claims require extraordinary evidence.
So if I tell someone the next day I got almost double the lumen output at under-driven currents, everyone will want to know how is that possible or simply ignore me, I would think that would be a fair attitude towards me.
So for example i was thinking to ask you this, maybe there is short somewhere, there are 4 current sense resistors on the driver, that simply means the current is being regulated (limited, as some people call it), but if the current sense resistors are bypassed or there is a short (ground input to ground output like commonly happens to short) then yeah the current output could be way higher becasue now current goes out stopped only the the entire resistance of the circuit. Maybe you can check if the current is still 8A on the output, you have checked it before.
Wow, that was one detailed explanation on your reflow technique! I think I just picked up a couple of keepers. Anyways, I don’t doubt your results. The next thing is to see if they can be replicated as you suggest. That my friend is a part of the scientific method. That would be a lot better than endless speculation. But through your endless modding Dale ( 
), you have inadvertently raised the bar for lumen output in such a small light. You should donate that light to the Southern California Museum of Flashlight History, which I’ll be trying to get off the ground . But seriously Dale, thanks for your constant modding adventures, and sharing those adventures with us all! You’re the best!
Thanks Ray-o-light, appreciate it.
Hikelight, if there were a negative to ground short I’d have no modes. Since I have all the modes, obviously there’s no short.
I could probably double check the sense resistors, see what value they are, but the more times you break down and reassemble a light the sloppier it gets. This one’s been apart 4 times now. Need a better reason to make it 5. This light was originally fitted with the XP-G2 S4 2B. On the box it says 6900 lumens. My meter was showing higher with a top cell, maybe they didn’t use the best high discharge cell when they read the 6900, I don’t know. But the light was in perfect working order with the XP-G2’s in it, swapping the emitters to XP-L HI made a TREMENDOUS difference.
You’re right, of course, that exceptional claims bring doubt and the desire to disprove. Human nature. That doesn’t change the fact that oncoming traffic on our road perceives this little light as a big 4x4 with lightbars fired up (my brother’s description, he was coming home when I was using this light Thursday night on my walk) {yes, I tend to shine a light down the road if I hear a car approaching over the rise, about 600 yds away before they crest the hill and can see the source}
You’re sounding like my Dad Hikelite, he’s about convinced they never sent a man to the moon at all, staged it, Hollywood photo’s and created rocks. Even my Dad is awed by this light, 86 year old cynic, grinning like the Cheshire cat with this little handful of light making a large room light up like daylight. He want’s to see it, push the button for himself, like I have some kind of trick up my sleeve or something. Shakes his head and says “… beats all I’ve ever seen!” My uncle came over to try to beat the old man in some Dominoes, Dad’s like, “go get that light of yours and show Jay…” all tickled and ready to see my Uncles reaction to this little thing.
I don’t actually understand where you picked up “cynical attitudes” or “desire to disprove” because it appears to me everybody’s desire is to prove, who in his right mind would ever want a M43 with 12k lumens not to be true? still, all things in this world can be scientifically explained and that’s what is being tried to accomplish.
btw, Vinh has listed M43 with XPL HI as ready to be sold so perhaps we’ll have some more info in these days!