At what current do you kill LEDs - which driver was used?

Thanks, looks like the 3-18V driver that IOS used to sell. A few others seems to sell it. LCK-LED being one of them..

Relic38 have resistor modded one (used for MT-G2). Feel free to share your experience about how it works in the high current driver thread (how far can you push it, and how fast does the overheat protection kick in and limit output once its modded).

If you kill the LED, or make it work nicely at 6A+, this is the thread. :)

As I think about it, I probably did buy it at IOS. Do you recall which amperage version they sold? I'm pretty sure 9 amps will not be survivable, but if it is a five or seven amp driver I will be in good shape.

I just checked my order history with IOS. I did buy it from them. It was listed as Boost/Buck 3-18V Triple XM-L Circuit Board". So it must be less than 9 amps. More likely much less.

Does Vinh mod his switches?

I have run this driver without problems with a XM-L2 U2 on a noctigon.

It’s a Ultrafire T-90 from wallbuys. I sold mine so I can’t check but it sure seems to be the same driver in the Lustefire U2.

I have used this driver to drive a single XM-L2 without problems and I can’t swear to it but I think it’s the same driver as in the Lustefire U2. I put a dry driver in my Lustefire and put this driver in a ZY-T08 to drive a MT-G2. I almost considered driving another XM-L2 de-domed with it, but I already have a ZY-T08 that is running a TR-3T6 driver with no problems, in fact I have two of them, one aspheric and one regular.

You guys must be using a resistor mod with your TR-3T6 and that is what is giving you the problems?

I’m waiting for the resin to dry on the MT-G2 so I don’t know yet how well it will run one MT-G2.

Im pretty sure it was rated 5A to single XM-l.

18650 wrote:

You guys must be using a resistor mod with your TR-3T6 and that is what is giving you the problems?

Yes, resister mods, but we're not sure its the driver giving problems as there are other differences that have not been isolated. Assuming reports of TN31's running at 6.5 amps without LED failures are accurate, then there is some cause other than over driving that has caused the LED failures we have so far documented in this thread. Since there are other variables, RaceR86 is trying to gather failure data to see if a pattern develops. So far there are 2 big differences between the successful TN31's and the dead LED's noted so far:

  • Heat sinking. The TN31 has a superior copper base and head.
  • Driver. The TN31's driver appears to regulate voltage and current.

I have fried my share of emitters and so far it has always been poor heatsinking that kills them or a short of some kind. Well that and bumping the wires on de-domed ones, that has taken out more than anything. But that is just being clumsy. I have also fried emitters by not allowing my epoxy to set up properly. I am using wakefield deltabond now and I let it set up overnight, amazing stuff.

I wonder how many problems people are having comes from not properly lapping and using cheaper stuff like arctic silver which from what I have read is not really all that great. I would think that if it’s running fine and then suddenly dies it’s either a short or the epoxy has loosened or dried and shrunk possibly from the heat.

That T90 I built was supposed to be running at close to 7amps and it had no problem on an aluminum base.

I did hook a couple of LEDs to my capacitive discharge spot welder and pumped 20,000 amps through them… they did not like it… at all… they got angry… very angry… explosively angry… bastards couldn’t take a joke…

LOL. That's crazy Texas. Hate to see what you do to an LED that it gets you mad.

18650 wrote:

. . . That T90 I built was supposed to be running at close to 7amps and it had no problem on an aluminum base.

Aluminum emitter base?

Do you know if it supposed to be 7 amps to the emitter?

Did you get a chance to measure current at the emitter?

No on a noctigon. My MM stinks so it’s not accurate enough to bother with. Relic put an XM-L2 in his too, so you could ask him, he might have done it.

Earlier in the week I put the ZY-T08 back together with the driver circuit from my Lustefire 3XU2. I modified the driver to have pretty much the same output as the modified TR-3T6. I measured 6,4A at the emitter. I have not done any of my typical extreme indoor tests, simply because the de-domed emitter I put into it had the best tint I have ever seen from a de-domed emitter, and if its going to die, I want to say that I was able to enjoy it.. I can report that I have about 1hour runtime on the light, most of which have been on high and its still going strong. If it fails, ill report back..

The driver, or similar can be found on ebay, search "sst90 driver". Example. Its often rated at 7A. I have seen reports of 7-9A stock on this driver. But I have also seen some ebay sellers in the past say 5A. Either way, output seems to depend resistor..

Im not going to draw any conclusions, but so far it seems to be doing a good job.

Input current is around 4A on NCR18650B that have seen some use (3,93v resting) . Rough numbers:

4A*3,75V (guesstimated voltage at load)*2=30W

6,4A * vf around 4=25,6W

Efficiency around 85% on high..

I got another of the 5amp LCK driver in, tested it on my original XM-L on 20mm Alu (Tough little buggar, withstanding all the abuse I’ve put it through in the last month or so). Kaput before I could get a reading on the DMM. I was using the xml to test direct drive capabilities of batteries earlier today, getting up to 4.9 amps on 4.13v Pany ncr18650B.

My DST should be up and running with XM-L2 U2-1A, today or tomorrow, probably without spring/wire mods just in case this driver is running a bit hot. :slight_smile:

Actually, to be truer to Mark Twain's quote - The report of my DST's emitter death was an exaggeration.

Details here.

Race man, do you recall if your 3T6 driver had the 2 matching components (in Bays Q1 and Q3) on the horizontal board below the sense resistors?

Yes, mine have the two matching components.

On a side note, do you or anyone know the peak output of this driver MF TR3T6. (say, if you run it with 3 XM-L2s in parallel).

Thanks for the info. I figured you did have those components since you were able to get above 6Amps without immediately frying your emitter.

I don't know the max current yet. I haven't tried beyond the 7amps in the DST. I can't recall any other cases of anyone trying.

The current version of the "3T6" driver (TR-0142B) was driven to 9amps with an MT-G2 by Moderator007. He said he stopped there because we was concerned for his cells. Although very similar, I consider the TR-0142B a completely different driver. Actually, it's marketed as a SST?0 driver. It's not xml friendly above 5.5 amps. It does not have those two components above and some different values on the existing components.

EDIT: Changed some wording above.

EDIT2: More changes. I just can't focus on typing tonight. Getting off line to go relax and mod.

Ok, thanks for info. If I ever try it to push the limits with 3 XM-L2s (in parallel) Ill report back in a more suitable thread. Not sure if that is safer though..

Edit:

Some writing on mine behind torroid: TR-0159D 2012-01-L1

TR-0159A3 (Written on lower board)

Sounds good. I really want to connect this driver to a different MCU, but I can't find a source for it anymore. So not worth the trouble.

Just wanted to point out that my emitter death should be deleted from the OP. I had assumed it had died, but it turned out an emitter wire unsoldered from the driver. I later did fry the emitter, but with a different driver that does not like xml's.

I intend to run an xml2 at 7 amps (maybe higher) again, but in a different light but with the same driver. And with Ouchy's MCU mod. I think my emitter survived 7 amps because I had it heat sinked quite well (Notigon reflowed to copper bar in a deep (at least 1") solid aluminum pill. I just need to heat sink the driver better next time.