dedoming and then fry I am using this driver sku# from fasttech 1278501 in this brass pill https://www.fasttech.com/p/1208600 and it fits well in this pill I use artic silver and when I cycle through the modes it hits high and after a couple of seconds you see smoke coming frome the emitter and then it turns blue then dedomed itself and then the emitter burns itself completely out ,I know the emitter is on an aluminum pill but if I used noctigon copper pads would the emitters hold up any better?are there any other options,are nd it’s not like this happened just once almost all the dd fet drivers and then this new driver I decided to try it went through 4 emitters today so iam open to suggestions
thank you all in advance
For high amps / fet drivers you need to use a noctigon, sinkpad or other DTP star.
Generally, you will need Cu MCPCB DTP (Noctigon or SinkPad) for >3A setup. You should have turned off your light immediately when you see tint shift in the beam in HIGH mode. LED is salvageable if it is push to the limit for brief period. It is a small investment ($1.5) for longer run with Cu MCPCB.
As others have mentioned, you should buy a direct copper pill. It’s a cheap part that will improve the life of your LED. A copper stay will also give higher output even in lights that don’t need one to prevent overheating.
However, I’m not sure just the lack of a direct copper pill is your problem. I’ve seen many lights that come stock with thin aluminum pills and do fine at 3 amps (such as the Nitecore EC11 and EA11)
You might have an issue with the arctic silver. Perhaps you are using too much or too little of it. Additionally, even with arctic silver you still need something holding the star to the pill, such as screws or the reflector (with an isolator disk to prevent the reflector from shorting the leadwires). If you’re testing the pill outside the light, the star might be lifting away from the pill due to lack of the reflector to hold it down. This would result in instant overheating.
devil is in details, at 2 A you should be ok with none dt star, if it is designed right, your problem can be due to poor reflow, poorly designed star, and not firm connection star\pill, it could be “hanging in the air” and no amount of paste will fix that, or your pill has very poor finish, and actual area that star\pill makes contact is very small.
one has to look carefully at these things, it can even be few things combined that contribute to your overheat
Did you make sure the star was flat on the back side, and the metal surface it contacts was flat? And use just a tiny bit of thermal goop to fill the microscopic irregularities?
It’s the same basic approach used for computer CPUs: cpu thermal contact flat how? - Google Search
Of course they run more power and heat than most LEDs, but that’s the same problem; approximating their methods is the way to fix it.
Details matter indeed :)
He uses and XP-L V5 2A (tint) and with this driver https://www.fasttech.com/p/1278501
lol, you are correct, details matter, i tought 2a was the current, lol. the driver is 5Amps, that is definatly overdriving, no wonder it frys leds, on none dt star.
so if I try say a xml2-u4 on a noctigon pad ,I should be okay ?
Are you seeing any lifting of the traces on the MCPCB? One of my M6’s had that happen and now I have 1 pad that will act exactly like you’re describing.
Folks, Long time no see... Since most of you are now aware of my testing procedure, I'll be brief here. The following is a test of the Cree XML2 on a standard aluminum 20mm star, along with another xml2 mounted on a copper sinkpad. (member vestureofblood was kind enough to send me out these for testing). [image] Graphs speak louder than words, so here they are. The first one is the XML2 on a standard star, followed by a comparison graph between a standard aluminum star and a copper sink…
Some good info there for you.
I’m not sure why they list it as a 5 amp driver when their own testing gave 4.2 amps. That 2 piece design may trip the thermal protection at 55° pretty quickly. May be best to pot the driver with a thermal transfer material.
Iam looking into it