I am the one who took the photo of the failed E21A in post #445. I can confirm that it failed because of a bad reflow. After I replaced it, it turboed just fine with a H10 using my custom firmware with reduced FET.
I have also used turbo with a H10 and stock emitters and they have been fine so far.
These LEDs are pretty unknown in the flashlight community. Maybe they can handle a lot of current, maybe they can’t. We don’t know. Only time and experimentation will tell.
For now, anybody who is worried should play it safe and avoid using high-discharge cells (like Vapcell H10).
Please let’s not ruin this light with a bigger, hamstrung driver.
It would be nice, but That driver was ruled out up front by vote that since it has to be a fair bit bigger due to the coils etc and the point of this light was tiny size. Since we can’t have our cake and eat it too, I’m ok with the lack of aa supoort so that we may have the shortest aa light on planet earth… Anyway,there’s lots of other lights out there for running aa’s.
Not that i prefer it, but I’m ok with a very brief flash of max turbo due to heating up quickly. I spend alot of time outside at night in the country and its nice have a super bright turbo available to be able to blast something further away with 1000+ lumens, identify it, then go back to a walking around level of 200 lumens.
Looks good. @Wurkkos, any chance of champagne anodization like the TS21? I think it would work really well.
Thanks for sharing your experience
> using my custom firmware with reduced FET
what output level does your custom firmware reduced Turbo to? (iirc you said 65% on reddit, do you know if that is equivalent to a ceiling of 98/150?). And or do you know how many lumens your reduced Turbo is?
> I have also used turbo with a H10 and stock emitters and they have been fine so far.
is that also with custom firmware, with Turbo reduced to 98/150? or at Full 150/150?
I don’t have any means of measuring the output, but I was able to measure the current at the tail. My custom firmware reduced FET to 67% (I used the PWM levels from the Fireflies PL47-219B firmware, it is also a FET + 1 x 7135 driver), and with a fully charged H10 it drew ~5A at the start of turbo and increased slightly as the emitters got warm.
The stock emitters on turbo with the H10 was using firmware with unrestricted FET, and I measured over 7A at the start of turbo and went up to around 8A as the emitters got warm.
Interesting.
When I first noticed the reduced output on the LED in mine, I took off the head and looked at it (this was the first time I had ever opened the head). To my eyes the output appeared uniform, but just much dimmer than the other LEDs. It actually might have only been producing 10% of the light of the other leds.
When the light was off, the top of the LED looked completely pristine. It actually looked better than the other 2 leds which had specs of dirt stuck to them.
After I opened the head and reflowed the LED it was still dimmer. But now it appears to be only one side of the LED is lighting up. That side is at full brightness, but the other does not light up at all. Overall this was an improvement before the reflow from maybe 10% to 50% brightness, but it is obviously still not working right.
Looking into the bezel of my light now it looks like the LED in your picture in post 445, except without any visible burn marks on the led surface.
You also mentioned that in this and your next post that your TS10 failing was caused by a failed reflow. Meaning it worked fine before the reflow.
My conclusion and recommendation:
Previously I thought there had been two TS10s that had suffered possible LED damage due to use of Vapcell H10 battery. That seemed like strong evidence that the LEDs couldnt’ take it.
However, based on Zumlin’s post it now appears that my TS10 is the only light that has suffered a possible malfunction in this way.
A sample size of one may be too little evidence to conclude that the H10 is the problem. The problem my light had might not be too much current to the LED. Perhaps it was a bad factory reflow or defective LED.
Perhaps no changes should be made to the driver and the light left as-is. Or maybe just change future versions of the driver to default to “simple” mode with a lower output. The user can then swap to “Advanced” mode and take the risk if they want more output.
thanks
67% of level 150 of 150 could be level 100 of 150
There are no reports of LED damage in the TS10 with the Wurkkos battery. Not even at 150/150 maximum. There is One report of LED damage from using H10 battery at 150/150 maximum.
some strategies to prevent possible damage from using a 10A battery at level 150/150
1. Use a 3A battery, like the one Wurkkos offers. :student:
2. Set Simple mode as default UI with ceiling of 100/150 :+1:
3. If one Must use a 10A battery, lower the ceiling by 67%, to 100/150, do not use 150/150 with a 10A battery, unless youre feeling lucky… :confounded:
4. change the mcpcb to allow the use of Nichia 519a :innocent:
or other combinations of ideas? 
Yeah, I did see that. Nice mods and work, by the way!
So to clarify, your stock emitters did not fail, it was the E21As that you flowed in for a mod, right? But Firelight here did have his stock emitter fail.
I wish Wurkkos had done a more fully thorough product testing before releasing this light to the public…like the TS21. I/we might have assumed that if they were releasing a Latticepower emitter with a FET driver, it would have been tested enough to show that it can handle the power fed to it with any of our available cells…else the driver would be changed to suit, as it seems they might be looking into now. If it’s really not suitable and people still want the turbo boost from a small cell/host, maybe just go with traditional robust emitters like the vast majority of lights use, right?
Comparing the Nichia we know that it isn’t capable of much current without special treatment (see previous threads from Clemence and Djozz and some others…a few years back…seems like 3A was pushing it to its edge). Not to knock any manufacturer but Latticepower (fka Latticebright) doesn’t have a great track record in this area compared to the majors, so it would be surprising if this emitter was capable of more than Nichia’s. But then again, I dunno…does this Latticepower use a ceramic base or is it the same plasticky material as the E21A?
Wurkkos has been doing a good job but these mistakes are disheartening……perhaps focused the wrong things and rushing products to market too fast. Overall I think the TS10 is a great little light, but maybe jumped the gun a bit. I don’t think it’s necessarily worthy of being an enthusiast’s light with these issues and the assembly choices that make modding and repairs difficult or destructive.
afaik, both the E21a and the CSP 2323 in the TS10 are the same size, fit the same footprint, and have the same 1.4A max pulse rating
afaict, the only problem is the 10A battery…
imo, if people stick to 3A battery, there is no problem
I’m not so sure.
10A seemed like the most likely culprit. But lots of people have used the TS10 with 10A batteries and so far mine has had the only known failure in this manner.
Zumlin confirmed that his similar looking failure was actually caused by a failed reflow after a modding attempt. Mine is the only known sample that had an led fail before any modding was done. I’m not sure we should be generalizing to problems with 10A cells, based on a sample size of 1. Also, I’m not convinced Wurkkos should be making heavy mods to nerf this light based on only a single failure.
While 10A still seems like the most likely cause, it is also possible that mine failed due to a defective LED from the factory or some other cause. It should also be noted the the other 2 leds in my light continue to operate perfectly even on H10, showing that the leds in question can actually take high current. At least most of the time.
I agree…we just don’t know much of anything right now, certainly not enough to push suggestions or cautions. And it’s only been a couple weeks or less since anyone had theirs in hand. If we see a significant amount of failures in the emitters then there will be more to look at. Maybe Wurkkos would be kind enough to send some of these emitters to one of our members that does good tests (if they’re willing to spend the time testing). Right now we’re all just guinea pigs in two different ways with this new model.
Firelight, I hope Wurkkos takes care of you with a replacement or something….they seem to be great about that just like Sofirn.
I tend to agree, one failure is not indicative of a serious problem. But the population of BLF users having the light may also be small in comparison to the total number sold. But with a higher likelihood of using the light on turbo… so who knows.
I had already set ceiling to 110/150. For a light this size, that is still very bright. I have pocket carried AA/14500 lights for lots of years that maxed out at 150 to 200 lumens and rarely had to use them on their highest mode.
I also put the battery that came with the light back into it to replace the H10. This really did not impact the light noticeably for the things I need it to do. Why not be careful?
I don’t think I will avoid using the H10 in the future, if nothing else for its capacity. But, I will likely avoid turbo when using it. Not a big deal, because I normally have larger lights with me that produce much more light and can run turbo for long periods without killing themselves or burning my hand. (yes the TS10 gets too hot for me to hold comfortably in Turbo after a very short time).
Of course this could change depending on what we see going forward. Could be it is a bit of panic over something that really isn’t a problem…
Sorry if I missed a posted comparison,but I’d be interested to know the max output difference between the H10 and a standard 3A cell. There really isn’t a heck of a lot of current options in 14500 cells it seems. I thought at one time Shockli marketed a 7A cell, but darned if I can find it. Guessing it would be a great choice for this application.
> I’d be interested to know the max output difference between the H10 and a standard 3A cell.
I tested the 10Amp H10 and 3Amp L10 batteries in a FWAA:
momentary maximum w 219c LEDs, on my meter:
1100 lumens vs 900 lumens
18% less output at momentary maximum, from the 3A cell
not a visually significant difference.
both experience thermal step down
no difference at the sustainable output levels, about 300 lumens or less, that do not trigger thermal step down
===
I think the choice of batteries comes down to whether or not someone wants to OverDrive the LEDs.
the E21a and the CSP 2323 in the TS10 are rated for 1.4A max pulse, each. Total is 4.2A
Using a 3A battery, prevents any overpowering of the LEDs, since the FET can only provide 3A total. Less than the max spec.
Using a 10A battery Allows the FET to overdrive the LEDs at 7 Amps or more… and that exceeds the rated maximum of the LEDs… by almost 3 Amps.
I think its fine to experiment with overdriving, and to use Turbo as the primary mode, with the H10 battery. If you are a modder, able to repair any damaged LEDs.
I also think it is good for people that do not do their own repairs, not to overdrive their LEDs, not to prioritize using Turbo, and instead prioritize lower, sustainable outputs, with the stock battery
Jon, both of them are listed at 1.4A DC continuous and 2.0A pulse, but of course we already know that the Nichia can handle a bit more (not much more, typical of most Nichia models). Same-ish power dissipation and Tj figures. Latticepower…history shows that they have not been very accurate in datasheet info (perhaps copying parameters to appear closest to what they are trying to emulate….the ol’ XML days…) They’ve come a long way, though, and nobody seems to have tested these in a bench test.
Using a “3A” battery here doesn’t ensure anything if we are sticking to these datasheet numbers, and it may also stress those cells if they are pushed to their max happy comfort range. I guess without a bench test of the emitters we need to be measuring amps at the emitter instead to see what they’re really getting. Haven’t undone mine yet and don’t really plan to but we should be sure the mcpcb contact with the host is also adequate, especially given the emitter footprint.
If we, collectively, had EVER stopped at datasheet max ratings, where would our hobby be today? Virtually all of our emitters are overdrive, mostly to great success and enjoyment. UV and colored emitters are the usual exception there.
Personally, for a $20 light with other flaws, I’m not too worried if it craps out. Haven’t used it a ton yet but I’ve run four cycles with Vapcell gold 3A and also H10…not yet measured current but will do so later this weekend…likely will keep running H10 and just see how it fares over time. All I can say is that we should slow down and observe, maybe, just as Wurkkos should have before releasing this model. 
Can you please describe your method (how many wires, how many hands, how to activate Turbo, etc.) when measuring the LED current on an Anduril light that uses both the flashlight's body (for LED current) and an inner tube (for control current) without the tail cap switch being installed?
Thanks
I have measured LED currents using a clamp meter and one wire on Anduril lights with a side switch using the method in the below video from forum member SammysHP:
But, I have not been able to accomplish something similar on a tail switch Anduril light even when using a second wire connected to the inner control tube and making and breaking this connection because it seemed like either I could not make and break at the proper timing to get the light to function or maybe I could not keep the inner tube in constant contact with the head of the light.
Did you ever see Matin’s video?
Thanks, I remember Martin but had not seen that video and method for using a short stiff section of wire as a temporary switch and it worked successfully for me compared to trying to use the other end of a longer wire.
Too bad Martin (forum member M4D M4X) has not been recently active on this forum.