Ok, I finished my pipe yesterday… seems I’m getting similar reading- (Not exact) to other members with other lights. You can put me on the list as a low priority.
I got you added to the list. I think Texaspyro is going to be up next, it sounds like MEM and n10sivern might need a little bit of time before they’re ready.
No problem at all… I’m in no hurry. I just started testing stuff last night. I’ve been going through reviews of stock lights I own and looking at others lumen figures and mine have so far been in the same 100s of lumens as other lights.
Tested the armytek predator pro at 480 lumens… Sky Ray king with the BLF driver and noctigons at 2250ish… K40m 2850ish at start up. BLF x6 NW at 899… I’ve got two 90 bends in the light and use the same cheap orange and gray light meter most use. I calibrated my light off a fenix ld15 with a factor of .229. The main things I can get to quite cooperate are headlamps yet. Also I have sanded the inside of the pipes to cut down on reflection and painted the outside with flat black.
Ok. I’ve got the lights in hand and have some bad news… the big light it pretty much unsuited to the task. It has some severe issues when running off of batteries. Also has some heating problems. (BTW, the light draws 2.793 amps, but also has a somewhat unstable (1.5KHz to 15 kHz) PWM signature even on high…).
It starts off around 350 lumens and flickers and giggles around for around 9 minutes then pops up to around 450 lumens and stabilizes. By then the case temperature is up to over 60 degrees C. Testing it with a bench supply, it starts off around 590 lumens and starts dropping as the light rapidly heats up. At 15 minutes it is down to 435 lumens and up to 70+ degrees C.
The running-on-battery issue is probably not due to some heat (or cold) related problems. I put a new battery into the hot light and it did the same thing… started on a lower/flickery output then popped up and stabilized. I think there is a switch/spring/contact issue. Or a driver issue with battery impedance. It seems to run fine off the bench supply, but is quite un-good on batteries.
I have not had a chance to play with the small light yet…
Ouch, that is some bad news. I knew it had heating up issues, but I thought that I sort of covered that with the runtime graphs. I did not know that the qlite revA driver (in which pwm issues on high setting are supposed to be fixed) was not stable.
I hope that the small light does what it should do. At least the driver is not more complicated than just two 7135 chips and a cap, nothing more. What also could help is that these drivers (from dx) have been there on the dx-shelf for years and still are with 'original' 7135 chips.
I'm not at all an electronics person, perhaps using 7135 based drivers is not the best way to make a constant current flashlight, but I know no better.
Yeah, one problem is the light takes 9 minutes to stabilize on batteries… by then, it is already rather warm and the lumens are down around 25% from max.
I think that a reference light should be large and drive the LEDs at a lower power level to minimize heating effects.
Still not sure what the problem is during those first nine minutes…
Ahh - Am I getting the lights next? Is the F3 worth it? Should I look at fixing it? I got a couple F3's - one is still stock, the other modded.
I think an XM-L2 or XP-L would have worked out better in the F3 - XP-G2's are notorious for high Vf, even more so when de-domed. I think we'd see less variation with cells, better consistency. Of course what's described now though is something else - agree, flaky connection/ground/etc.
Agree bout bigger light, lower amps. But of course trying to save int'l shipping costs, and also have a higher powered light to measure - more lumens, more throw - tough decisions...
I think a 7135 based driver is a good way to go, specially at a level in the 3A range, no higher. Maybe an LD-1 or LD-2 is also a good choice to be PWM-less.
The idea was that I tried to build the F3 without any flaky connections, I swapped the switch as well because I know that the stock F3 switch causes trouble, and I would be surprised if this appears to be a connections problem. The idea behind the XP-G2 was pure the amount of throw that was required.
The second light was more what I had in mind originally. I agree that the best would be a UF-T20 or so, with a XP-L Hi at just 700mA, but this small Cofly zoomie can be packed narrower than 3.5cm and therefore I can send it as a 'letterbox package', saving 75% off the shipping costs. I hope that it suffices better than the F3.
No problem at all - just wondering what we do now, or what the next step is - sorry, was trying to help.
And all I intended was trying to explain what I had done, and all help is appreciated :-)
I disassembled the tail cap assembly on the big light, applied some contact cleaning rainbow flavored unicorn tears, adjusted the side contact of the switch (that touches the body of the tailcap, sacrificed a couple of virgin unicorns, etc. Light seems to be behaving better now. I need to run it and see if the instabilities come back.
Also did a quick test run on the small light. Starts out at 115 lumens with CCT at 4153K, after 25 minutes is down to 105 lumens, 4183K, 46 degrees C… much better behavior for a calibration light.
I would like to be added to the list please.
Despite the limited quality of the flashlights sent around, this is turning out to be a more useful undertaking than I thought. The differences between me, KKW and tp appear not a few percent, we're talking tens of percents! I will one of these days start a table in the OP with the data from the different participants.
Agreed, I’m glad you decided to build and send the second light especially. I had a feeling my meter probably read 15% low, and was surprised that it came back as 20% low. My lumens calibration was within about 6% of a djozz lumen, so I feel like I did a decent job with my initial calibration numbers from the lights I used.
Once again, thanks djozz for the lights for this, I hope you have some use for the battery and charger when everything gets back to you.
Added
I tried another test of the big light today and it is back to its evil ways, but a little different. It started out around 420 lumens and dropped to around 350 lumens after 10 minutes as it warmed up. Then it started going unstable.
Running on a battery, it does not get as hot as when run off of a bench supply. Probably due to the extra thermal mass the battery adds along with the voltage sag of the battery. It only gets to around 62 degrees C running off of a battery vs 70+ C on a 4.2V bench supply.
The Cofty light has almost exactly a 2:1 lumen range on flood to throw.
I take it that you used as the throw position the sharply imaged die? (the head fully out brings it beyond the sharp image). The difference between my 0.478 and your 0.5 (4.5% variation), that implies that one of the spheres integrates worse than the other. Hmm 
, I have not done that yet, but I will check my 'sphere collection' at home for this variation (I made a similar Cofly for own use as well).
Here's my specs of the small Cofly again:
Specifications:
*host=Cofly KX-H10
*led=dedomed XP-E2 R3 1D0 on 16mm Noctigon
*driver=2×7135 plain (no protections, no modes) linear driver, 710mA measured
*output flood=136 djozzlumen, measured between 8 and 10 minutes. (ftr, at 30 seconds, 143 lumen was measured)
*throw/flood ratio=0.478 , where the throw position of the lens is set at a sharply imaged die at 5 meters (a bit closer or further does not matter much)
*throw=30.1 kcd, 30.4 kcd resp. (two measurements on two different batteries of the same type), measured again between 8 and 10 minutes, at 5 meter distance, searching for the highest reading while moving the hotspot over the luxmeter detector.
Handling of the flashlight:
*if not needed, do not open the head of the light
*do not place the flashlight vertically on the lens, it will scratch because it sticks out a bit ( there is already a tiny scratchmark in the middle)
*avoid smudging the lens, it is plastic and will scratch easily, also when cleaned, a scratched lens will influence the throw number.
*this flashlight should be waterproof but to avoid having to clean the lens (introducing extra scratchmarks), do not use in wet conditions.
Poll. So how many bends do each the participants have in their pipes? Did you ever have one with less or more bends. I have two now. Adding a third would be easy, but are there significant improvements?
This is a picture of mine. Its made from 90mm down pipe as it is a lot cheaper than 100 mm pipe. I may have to build another with 100mm as some of my lights dont fit.
Tonight I measured the throw/flood ratio of a copy (same host, same build) of the small Cofly zoomie that I made for own use. The ratio differs a bit from the one that is being sent around (the flood position of the lens is probably a tiny bit different), but the point is that I measured it in all of my 4 integrating spheres, that vary a lot in size, design, coating of the inside and type of luxmeter used (three different ones), to see if the flood/throw ratio of this flashlight between them is comparable (which would be an indication of comparable integration). I measured 4 times per sphere to see the influence of slight varying position of the flashlight in the entrance hole, and slightly varying lens position every time it is slided back and forth. These are the results:
Sphere #1 ratio measurement varies from 0.449 to 0.455
Sphere #2 ratio measurement varies from 0.449 to 0.451
Sphere #3 ratio measurement varies from 0.450 to 0.457
Sphere #4 ratio measurement varies from 0447 to 0.451
All 4 look pretty indistinguishable to me. So either all 4 spheres have bad integration but then equally bad, or equally good. I like to think that it is more probable that the integration is good in all 4.
I hope to see these flashlights back after the world tour to see if I still measure the same numbers.