Strange Thrunite TN31 problem

Hi!

I hope I post in the right sub-forum…

Well, a friend of mine bought a Thrunite TN31, second hand so no warranty. After some time (several months), the led died suddently. He had the led replaced (by an XM-L2) by a local modder, and the light was working fine again. Then after ~2 weeks the led fried again. He sent me the light so I can replaced the led and have a look inside.

I replaced the led, which was strangely fried (no change in the color of the led, no sign of overheating or anything). Well, so I replaced the led, took extra care of the thermal path (no kapton tape under the led!), sanded the star and the pill, used high quality thermal paste, etc etc… I think I did everything by the book, and the light was working fine again.

I took some measurements, it was pulling 1,63A 12V, which is quite normal for a stock TN31, and ~1,23A after step down (still 12V), which I assume is also normal for a stock TN31. I tested the light in several conditions, noticed it was working perfectly fine, does not seem to overheat, etc…and sent the light back to my friend.

This was like three weeks ago…and this morning he told me the led fried again…

Now I am clueless…why? What happened to the fried leds? They doesn’t look overheated or anything. I’ve seen “dead” and cooked leds, and in the TN31 it doesn’t look like anything I’ve ever seen, it just looks like a normal working XM-L2.

So, this is very embarrassing…the TN31 driver is completely stock (I resisted the urge to do a small resistor mod on it), the thermal path was good, why did the led fried?

My guess is, there is something with the driver which fried the leds…like reverse polarity to the led…don’t know if it’s possible, but it is the only explanation I can think of right now.

What do you think? A way to fix it? I’m pretty sure if I replace the led again it will work again for some time…

Thanks in advance for any idea or suggestion!

If you exceed about 8A or so, you could fry a LED with nothing obvious looking, but those really small wires to the phosphor will be disintegrated. Did you look with a magnifier at the dead LED's? I just did that to a XM-L2 LED last night in fact ().

Maybe there's some sort of current spike occurring, not sure what reason though, but faulty driver or a short. With a short I've killed LED's and they looked ok, think (not sure) the small wires were still in tact.

Not sure if ThruNite will provide a replacement driver, but if you can't see any obvious shorts, you can try contacting them. I would replace the LED wires also just in case, use 18 or 20 gauge wires too.

Thanks for the information. The led wires are intact, but maybe the driver has a spike problem…I’ll try to contact Thrunite for a replacement…

In the meantime, do you still think it is worth replacing the led once again? And the wires too, why not.

Well I'd examine those wires and their soldered pts of contact very carefully, and if you find something/anything suspect - great, replace them, new LED and go. If you really see nothing -- dunno if it's worth the risk of another frustrating failure. Also, I'd examine the driver very carefully, clean it up thoroughly with isopropyl alcohol, use a nylon brush to get it sparkly. Also check the top half of the driver to be sure nothing sticks up. The cover going over the driver is metal, so possible short may occur there. Check all clearances around the spring, everything. A weakening spring could be it may move slightly under normal usage, possibly short out, etc.

It may be worth the risk after doing all the above, even if you didn't notice anything specifically. Could be a solder wisk you didn't notice, minute particle(s), partial conductive materials (maybe oils, grease, old flux, etc.).

My experiences with these lights, though they are excellent quality, is that the quality control and manufacturing/assembly is spotty, and do have manny quality variations from piece to piece. I've done work on maybe a dozen or so of these lights. They do have issues all over, like the battery carrier, inconsistent performance of the reflector, etc. Might be worthwhile doing a review of the carrier as well, and tailcap area - may be hard to remove though.

You sure sound like you know what you're doing and your hunch sounds right. If you don't have an oscilloscope, you could hook up a volt meter across a sacrificial LED (or a dummy load made of diodes) and look for voltage spikes while operating it and changing modes.

Are you sure the bond wires in the LED's are still intact? Maybe try to take a macro shot and magnify on computer screen to make sure.

Thank you for your answers. I carefully examined the both leds (the one I put and the one I replaced) with a Solarforce L2CL1 (:D) as a magnifier, and I’m pretty sure both leds have their bonds intact.
So we’re back on the driver issue. I’ve cleaned it a little when I replaceed the led, and found nothing unusual, but I’ll double check that point. And replace the wires, can’t harm!

I’ve already checked the battery carrier the first time, tightened the srews and tested it through an array of 3 leds, I think it is safe to say it is ok.

Well well well, so I’d try to get a hand on an oscilloscope, my father has one, but it is so old it has difficulties catching PWM in computer fans PWM signal so…I heard about soundcard oscilloscope, maybe have a look at it.

I’ll keep you updated :slight_smile:

Current spikes in a faulty driver can do that, especially when u turn on and off I believe. As other have said u better proceed with oscilloscope, monitoring voltage and amperage.

Hmm which thermal paste are you using between the PCB and the base? You might want to replace it with a better quality one.

I used Thermalright Chillfactor III. For computer use, it is one of my favourites, along with Prolimatech PK3 :slight_smile:
I don’t like AS5, it needs a curing time, dries eventually, and is far from the best when contact pressure isn’t optimal (which is almost always the case in flashlight use)….well, I also use Fujik thermal adhesive because it’s convenient and cheap, when I have no other choice (Arctic adhesive is quite hard to find over there), not often!

I took extra care during the application, I put a thin layer of paste with a spatula…most of the time I do this for my cpu coolers, and it is fine. I was a bit unsure on how to do it, since that even with the two screws the pressure applied if far from the pressure applied by a proper monting system of a cpu cooler, but considering the amount of dried thermal paste I found under the star when I first opened it, I thought it would be ok :wink:

I think CFIII is appropriate, but if you disagree I don’t mind trying with another paste next time, I got plenty of different brands in my drawer :slight_smile:
For the moment I’m waiting for an answer from Thrunite regarding a replacement driver… and I’m waiting to get a hand on an oscilloscope…

TC3 should be more than adequate, so I doubt that’s the issue.

The LED I’m assuming must have been reflowed well too, perhaps the driver’s the culprit.

Let us know how this turns out!

Well, I tried my best with the led too :slight_smile:

I’ll keep you updated as soon as I have anything interesting to add!

To the LED experts out there. Are there other typical failure points in xml2's other than the bonding wires? Not counting physical damage to dome and phosphor layer? Just the parts that create photons.

This case is interesting because 2 xml2's have failed. Both appear to have intact bond wires. I take it you tested the failed emitters outside of the light. Did you try to reflow them to a different base and test to see if they still don't work? I'm just wondering what failed in them.

In fact I tried them on the Thrunite pcb and then directly bare led with a lab power supply, the first one appears completely dead, the second one glows very very low…I should have mentionned that earlier, sorry. I too wonder why they failed…

When do the leds die? Do you just turn the torch on and find its failed or did they fail in use or when adjusting the selector ring?

I’m not the one using the light, but I understood that one day the light doesn’t work, so I’d say the led fail when the light is turned on. But I can’t be 100% sure on that.

I has an XM-L die like that when I accidentally direct drived it with a 3S pack; I couldn’t see any signs of physical damage and the bond wires were intact.

Ok, so it could definitely lead to a driver problem… Hmmm…
Waiting for Thrunite to answer (maybe…)…or do you have an idea for a replacement driver?

I got an answer from Thrunite asking if I wanted the driver only or the driver + the led…now I’m waiting for the price, will keep you updated…hope it won’t be too expensive!

Same here. Best of luck.

You may want to contact Supfire Supbeam. Sirius9 recommended here in post 40 to contact Supfire Supbeam for drivers for a ThruNite. The Supfire Supbeam K40 uses and almost identical driver to the TN31.

Good idea, thanks.
For the moment Thrunite told me I’d just have to pay the shipping…sounds good for now!